/root/doris/be/src/common/config.cpp
Line | Count | Source (jump to first uncovered line) |
1 | | // Licensed to the Apache Software Foundation (ASF) under one |
2 | | // or more contributor license agreements. See the NOTICE file |
3 | | // distributed with this work for additional information |
4 | | // regarding copyright ownership. The ASF licenses this file |
5 | | // to you under the Apache License, Version 2.0 (the |
6 | | // "License"); you may not use this file except in compliance |
7 | | // with the License. You may obtain a copy of the License at |
8 | | // |
9 | | // http://www.apache.org/licenses/LICENSE-2.0 |
10 | | // |
11 | | // Unless required by applicable law or agreed to in writing, |
12 | | // software distributed under the License is distributed on an |
13 | | // "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY |
14 | | // KIND, either express or implied. See the License for the |
15 | | // specific language governing permissions and limitations |
16 | | // under the License. |
17 | | |
18 | | #include <fmt/core.h> |
19 | | #include <stdint.h> |
20 | | |
21 | | #include <algorithm> |
22 | | #include <cctype> |
23 | | // IWYU pragma: no_include <bthread/errno.h> |
24 | | #include <lz4/lz4hc.h> |
25 | | |
26 | | #include <cerrno> // IWYU pragma: keep |
27 | | #include <cstdlib> |
28 | | #include <cstring> |
29 | | #include <fstream> // IWYU pragma: keep |
30 | | #include <functional> |
31 | | #include <iostream> |
32 | | #include <map> |
33 | | #include <memory> |
34 | | #include <mutex> |
35 | | #include <random> |
36 | | #include <string> |
37 | | #include <utility> |
38 | | #include <vector> |
39 | | |
40 | | #include "cloud/config.h" |
41 | | #include "common/config.h" |
42 | | #include "common/logging.h" |
43 | | #include "common/status.h" |
44 | | #include "io/fs/file_writer.h" |
45 | | #include "io/fs/local_file_system.h" |
46 | | #include "util/cpu_info.h" |
47 | | |
48 | | namespace doris::config { |
49 | | |
50 | | // Dir of custom config file |
51 | | DEFINE_String(custom_config_dir, "${DORIS_HOME}/conf"); |
52 | | |
53 | | // Dir of jdbc drivers |
54 | | DEFINE_String(jdbc_drivers_dir, "${DORIS_HOME}/jdbc_drivers"); |
55 | | |
56 | | // cluster id |
57 | | DEFINE_Int32(cluster_id, "-1"); |
58 | | // port on which BackendService is exported |
59 | | DEFINE_Int32(be_port, "9060"); |
60 | | |
61 | | // port for brpc |
62 | | DEFINE_Int32(brpc_port, "8060"); |
63 | | |
64 | | DEFINE_Int32(arrow_flight_sql_port, "-1"); |
65 | | |
66 | | // If the external client cannot directly access priority_networks, set public_host to be accessible |
67 | | // to external client. |
68 | | // There are usually two usage scenarios: |
69 | | // 1. in production environment, it is often inconvenient to expose Doris BE nodes to the external network. |
70 | | // However, a reverse proxy (such as Nginx) can be added to all Doris BE nodes, and the external client will be |
71 | | // randomly routed to a Doris BE node when connecting to Nginx. set public_host to the host of Nginx. |
72 | | // 2. if priority_networks is an internal network IP, and BE node has its own independent external IP, |
73 | | // but Doris currently does not support modifying priority_networks, setting public_host to the real external IP. |
74 | | DEFINE_mString(public_host, ""); |
75 | | |
76 | | // If the BE node is connected to the external network through a reverse proxy like Nginx |
77 | | // and need to use Arrow Flight SQL, should add a server in Nginx to reverse proxy |
78 | | // `Nginx:arrow_flight_sql_proxy_port` to `BE_priority_networks:arrow_flight_sql_port`. For example: |
79 | | // upstream arrowflight { |
80 | | // server 10.16.10.8:8069; |
81 | | // server 10.16.10.8:8068; |
82 | | //} |
83 | | // server { |
84 | | // listen 8167 http2; |
85 | | // listen [::]:8167 http2; |
86 | | // server_name doris.arrowflight.com; |
87 | | // } |
88 | | DEFINE_Int32(arrow_flight_sql_proxy_port, "-1"); |
89 | | |
90 | | // the number of bthreads for brpc, the default value is set to -1, |
91 | | // which means the number of bthreads is #cpu-cores |
92 | | DEFINE_Int32(brpc_num_threads, "256"); |
93 | | // the time of brpc server keep idle connection, setting this value too small may cause rpc between backends to fail, |
94 | | // the default value is set to -1, which means never close idle connection. |
95 | | DEFINE_Int32(brpc_idle_timeout_sec, "-1"); |
96 | | |
97 | | // Declare a selection strategy for those servers have many ips. |
98 | | // Note that there should at most one ip match this list. |
99 | | // this is a list in semicolon-delimited format, in CIDR notation, e.g. 10.10.10.0/24 |
100 | | // If no ip match this rule, will choose one randomly. |
101 | | DEFINE_String(priority_networks, ""); |
102 | | |
103 | | // memory mode |
104 | | // performance or compact |
105 | | DEFINE_String(memory_mode, "moderate"); |
106 | | |
107 | | DEFINE_mBool(enable_use_cgroup_memory_info, "true"); |
108 | | |
109 | | // process memory limit specified as number of bytes |
110 | | // ('<int>[bB]?'), megabytes ('<float>[mM]'), gigabytes ('<float>[gG]'), |
111 | | // or percentage of the physical memory ('<int>%'). |
112 | | // defaults to bytes if no unit is given" |
113 | | // must larger than 0. and if larger than physical memory size, |
114 | | // it will be set to physical memory size. |
115 | | DEFINE_String(mem_limit, "90%"); |
116 | | |
117 | | // Soft memory limit as a fraction of hard memory limit. |
118 | | DEFINE_Double(soft_mem_limit_frac, "0.9"); |
119 | | |
120 | | // Cache capacity reduce mem limit as a fraction of soft mem limit. |
121 | | DEFINE_mDouble(cache_capacity_reduce_mem_limit_frac, "0.6"); |
122 | | |
123 | | // Schema change memory limit as a fraction of soft memory limit. |
124 | | DEFINE_Double(schema_change_mem_limit_frac, "0.6"); |
125 | | |
126 | | // Many modern allocators (for example, tcmalloc) do not do a mremap for |
127 | | // realloc, even in case of large enough chunks of memory. Although this allows |
128 | | // you to increase performance and reduce memory consumption during realloc. |
129 | | // To fix this, we do mremap manually if the chunk of memory is large enough. |
130 | | // |
131 | | // The threshold (128 MB, 128 * (1ULL << 20)) is chosen quite large, since changing the address |
132 | | // space is very slow, especially in the case of a large number of threads. We |
133 | | // expect that the set of operations mmap/something to do/mremap can only be |
134 | | // performed about 1000 times per second. |
135 | | // |
136 | | // P.S. This is also required, because tcmalloc can not allocate a chunk of |
137 | | // memory greater than 16 GB. |
138 | | DEFINE_mInt64(mmap_threshold, "134217728"); // bytes |
139 | | |
140 | | // When hash table capacity is greater than 2^double_grow_degree(default 2G), grow when 75% of the capacity is satisfied. |
141 | | // Increase can reduce the number of hash table resize, but may waste more memory. |
142 | | DEFINE_mInt32(hash_table_double_grow_degree, "31"); |
143 | | |
144 | | DEFINE_mInt32(max_fill_rate, "2"); |
145 | | |
146 | | DEFINE_mInt32(double_resize_threshold, "23"); |
147 | | |
148 | | // The maximum low water mark of the system `/proc/meminfo/MemAvailable`, Unit byte, default -1. |
149 | | // if it is -1, then low water mark = min(MemTotal - MemLimit, MemTotal * 5%), which is 3.2G on a 64G machine. |
150 | | // Turn up max. more memory buffers will be reserved for Memory GC. |
151 | | // Turn down max. will use as much memory as possible. |
152 | | // note that: `max_` prefix should be removed, but keep it for compatibility. |
153 | | DEFINE_Int64(max_sys_mem_available_low_water_mark_bytes, "-1"); |
154 | | |
155 | | DEFINE_Int64(memtable_limiter_reserved_memory_bytes, "838860800"); |
156 | | |
157 | | // The size of the memory that gc wants to release each time, as a percentage of the mem limit. |
158 | | DEFINE_mString(process_minor_gc_size, "5%"); |
159 | | DEFINE_mString(process_full_gc_size, "10%"); |
160 | | |
161 | | // If true, when the process does not exceed the soft mem limit, the query memory will not be limited; |
162 | | // when the process memory exceeds the soft mem limit, the query with the largest ratio between the currently |
163 | | // used memory and the exec_mem_limit will be canceled. |
164 | | // If false, cancel query when the memory used exceeds exec_mem_limit, same as before. |
165 | | DEFINE_mBool(enable_query_memory_overcommit, "true"); |
166 | | |
167 | | DEFINE_mBool(disable_memory_gc, "false"); |
168 | | |
169 | | DEFINE_mBool(enable_stacktrace, "true"); |
170 | | |
171 | | DEFINE_mInt64(stacktrace_in_alloc_large_memory_bytes, "2147483648"); |
172 | | |
173 | | DEFINE_mInt64(crash_in_alloc_large_memory_bytes, "-1"); |
174 | | |
175 | | // The actual meaning of this parameter is `debug_memory`. |
176 | | // 1. crash in memory tracker inaccurate, if memory tracker value is inaccurate, BE will crash. |
177 | | // usually used in test environments, default value is false. |
178 | | // 2. print more memory logs. |
179 | | DEFINE_mBool(crash_in_memory_tracker_inaccurate, "false"); |
180 | | |
181 | | // default is true. if any memory tracking in Orphan mem tracker will report error. |
182 | | // !! not modify the default value of this conf!! otherwise memory errors cannot be detected in time. |
183 | | // allocator free memory not need to check, because when the thread memory tracker label is Orphan, |
184 | | // use the tracker saved in Allocator. |
185 | | DEFINE_mBool(enable_memory_orphan_check, "true"); |
186 | | |
187 | | // The maximum time a thread waits for full GC. Currently only query will wait for full gc. |
188 | | DEFINE_mInt32(thread_wait_gc_max_milliseconds, "1000"); |
189 | | |
190 | | DEFINE_mInt64(pre_serialize_keys_limit_bytes, "16777216"); |
191 | | |
192 | | // the port heartbeat service used |
193 | | DEFINE_Int32(heartbeat_service_port, "9050"); |
194 | | // the count of heart beat service |
195 | | DEFINE_Int32(heartbeat_service_thread_count, "1"); |
196 | | // the count of thread to create table |
197 | | DEFINE_Int32(create_tablet_worker_count, "3"); |
198 | | // the count of thread to drop table |
199 | | DEFINE_Int32(drop_tablet_worker_count, "3"); |
200 | | // the count of thread to batch load |
201 | | DEFINE_Int32(push_worker_count_normal_priority, "3"); |
202 | | // the count of thread to high priority batch load |
203 | | DEFINE_Int32(push_worker_count_high_priority, "3"); |
204 | | // the count of thread to publish version |
205 | | DEFINE_Int32(publish_version_worker_count, "8"); |
206 | | // the count of tablet thread to publish version |
207 | | DEFINE_Int32(tablet_publish_txn_max_thread, "32"); |
208 | | // the timeout of EnginPublishVersionTask |
209 | | DEFINE_Int32(publish_version_task_timeout_s, "8"); |
210 | | // the count of thread to calc delete bitmap |
211 | | DEFINE_Int32(calc_delete_bitmap_max_thread, "32"); |
212 | | // the count of thread to calc delete bitmap worker, only used for cloud |
213 | | DEFINE_Int32(calc_delete_bitmap_worker_count, "8"); |
214 | | // the count of thread to calc tablet delete bitmap task, only used for cloud |
215 | | DEFINE_Int32(calc_tablet_delete_bitmap_task_max_thread, "32"); |
216 | | // the count of thread to clear transaction task |
217 | | DEFINE_Int32(clear_transaction_task_worker_count, "1"); |
218 | | // the count of thread to delete |
219 | | DEFINE_Int32(delete_worker_count, "3"); |
220 | | // the count of thread to alter table |
221 | | DEFINE_Int32(alter_tablet_worker_count, "3"); |
222 | | // the count of thread to alter index |
223 | | DEFINE_Int32(alter_index_worker_count, "3"); |
224 | | // the count of thread to clone |
225 | | DEFINE_Int32(clone_worker_count, "3"); |
226 | | // the count of thread to clone |
227 | | DEFINE_Int32(storage_medium_migrate_count, "1"); |
228 | | // the count of thread to check consistency |
229 | | DEFINE_Int32(check_consistency_worker_count, "1"); |
230 | | // the count of thread to upload |
231 | | DEFINE_Int32(upload_worker_count, "1"); |
232 | | // the count of thread to download |
233 | | DEFINE_Int32(download_worker_count, "1"); |
234 | | // the count of thread to make snapshot |
235 | | DEFINE_Int32(make_snapshot_worker_count, "5"); |
236 | | // the count of thread to release snapshot |
237 | | DEFINE_Int32(release_snapshot_worker_count, "5"); |
238 | | // report random wait a little time to avoid FE receiving multiple be reports at the same time. |
239 | | // do not set it to false for production environment |
240 | | DEFINE_mBool(report_random_wait, "true"); |
241 | | // the interval time(seconds) for agent report tasks signature to FE |
242 | | DEFINE_mInt32(report_task_interval_seconds, "10"); |
243 | | // the interval time(seconds) for agent report disk state to FE |
244 | | DEFINE_mInt32(report_disk_state_interval_seconds, "60"); |
245 | | // the interval time(seconds) for agent report olap table to FE |
246 | | DEFINE_mInt32(report_tablet_interval_seconds, "60"); |
247 | | // the max download speed(KB/s) |
248 | | DEFINE_mInt32(max_download_speed_kbps, "50000"); |
249 | | // download low speed limit(KB/s) |
250 | | DEFINE_mInt32(download_low_speed_limit_kbps, "50"); |
251 | | // download low speed time(seconds) |
252 | | DEFINE_mInt32(download_low_speed_time, "300"); |
253 | | |
254 | | DEFINE_String(sys_log_dir, ""); |
255 | | DEFINE_String(user_function_dir, "${DORIS_HOME}/lib/udf"); |
256 | | // INFO, WARNING, ERROR, FATAL |
257 | | DEFINE_mString(sys_log_level, "INFO"); |
258 | | // TIME-DAY, TIME-HOUR, SIZE-MB-nnn |
259 | | DEFINE_String(sys_log_roll_mode, "SIZE-MB-1024"); |
260 | | // log roll num |
261 | | DEFINE_Int32(sys_log_roll_num, "10"); |
262 | | // verbose log |
263 | | DEFINE_Strings(sys_log_verbose_modules, ""); |
264 | | // verbose log level |
265 | | DEFINE_Int32(sys_log_verbose_level, "10"); |
266 | | // verbose log FLAGS_v |
267 | | DEFINE_Int32(sys_log_verbose_flags_v, "-1"); |
268 | | // log buffer level |
269 | | DEFINE_String(log_buffer_level, ""); |
270 | | // log enable custom date time format |
271 | | DEFINE_Bool(sys_log_enable_custom_date_time_format, "false"); |
272 | | // log custom date time format (https://en.cppreference.com/w/cpp/io/manip/put_time) |
273 | | DEFINE_String(sys_log_custom_date_time_format, "%Y-%m-%d %H:%M:%S"); |
274 | | // log custom date time milliseconds format (fmt::format) |
275 | | DEFINE_String(sys_log_custom_date_time_ms_format, ",{:03d}"); |
276 | | |
277 | | // number of threads available to serve backend execution requests |
278 | | DEFINE_Int32(be_service_threads, "64"); |
279 | | |
280 | | // The pipeline task has a high concurrency, therefore reducing its report frequency |
281 | | DEFINE_mInt32(pipeline_status_report_interval, "10"); |
282 | | // number of scanner thread pool size for olap table |
283 | | // and the min thread num of remote scanner thread pool |
284 | | DEFINE_Int32(doris_scanner_thread_pool_thread_num, "-1"); |
285 | | DEFINE_Validator(doris_scanner_thread_pool_thread_num, [](const int config) -> bool { |
286 | | if (config == -1) { |
287 | | CpuInfo::init(); |
288 | | doris_scanner_thread_pool_thread_num = std::max(48, CpuInfo::num_cores() * 2); |
289 | | } |
290 | | return true; |
291 | | }); |
292 | | DEFINE_Int32(doris_scanner_min_thread_pool_thread_num, "8"); |
293 | | DEFINE_Int32(remote_split_source_batch_size, "10240"); |
294 | | DEFINE_Int32(doris_max_remote_scanner_thread_pool_thread_num, "-1"); |
295 | | // number of olap scanner thread pool queue size |
296 | | DEFINE_Int32(doris_scanner_thread_pool_queue_size, "102400"); |
297 | | // default thrift client connect timeout(in seconds) |
298 | | DEFINE_mInt32(thrift_connect_timeout_seconds, "3"); |
299 | | DEFINE_mInt32(fetch_rpc_timeout_seconds, "30"); |
300 | | |
301 | | // default thrift client retry interval (in milliseconds) |
302 | | DEFINE_mInt64(thrift_client_retry_interval_ms, "1000"); |
303 | | // max message size of thrift request |
304 | | // default: 100 * 1024 * 1024 |
305 | | DEFINE_mInt64(thrift_max_message_size, "104857600"); |
306 | | // max bytes number for single scan range, used in segmentv2 |
307 | | DEFINE_mInt32(doris_scan_range_max_mb, "1024"); |
308 | | // single read execute fragment row number |
309 | | DEFINE_mInt32(doris_scanner_row_num, "16384"); |
310 | | // single read execute fragment row bytes |
311 | | DEFINE_mInt32(doris_scanner_row_bytes, "10485760"); |
312 | | // single read execute fragment max run time millseconds |
313 | | DEFINE_mInt32(doris_scanner_max_run_time_ms, "1000"); |
314 | | // (Advanced) Maximum size of per-query receive-side buffer |
315 | | DEFINE_mInt32(exchg_node_buffer_size_bytes, "20485760"); |
316 | | DEFINE_mInt32(exchg_buffer_queue_capacity_factor, "64"); |
317 | | |
318 | | // memory_limitation_per_thread_for_schema_change_bytes unit bytes |
319 | | DEFINE_mInt64(memory_limitation_per_thread_for_schema_change_bytes, "2147483648"); |
320 | | |
321 | | DEFINE_mInt32(cache_prune_interval_sec, "10"); |
322 | | DEFINE_mInt32(cache_periodic_prune_stale_sweep_sec, "60"); |
323 | | // the clean interval of tablet lookup cache |
324 | | DEFINE_mInt32(tablet_lookup_cache_stale_sweep_time_sec, "30"); |
325 | | DEFINE_mInt32(point_query_row_cache_stale_sweep_time_sec, "300"); |
326 | | DEFINE_mInt32(disk_stat_monitor_interval, "5"); |
327 | | DEFINE_mInt32(unused_rowset_monitor_interval, "30"); |
328 | | DEFINE_mInt32(quering_rowsets_evict_interval, "30"); |
329 | | DEFINE_String(storage_root_path, "${DORIS_HOME}/storage"); |
330 | | DEFINE_mString(broken_storage_path, ""); |
331 | | |
332 | | // Config is used to check incompatible old format hdr_ format |
333 | | // whether doris uses strict way. When config is true, process will log fatal |
334 | | // and exit. When config is false, process will only log warning. |
335 | | DEFINE_Bool(storage_strict_check_incompatible_old_format, "true"); |
336 | | |
337 | | // BE process will exit if the percentage of error disk reach this value. |
338 | | DEFINE_mInt32(max_percentage_of_error_disk, "100"); |
339 | | DEFINE_mInt32(default_num_rows_per_column_file_block, "1024"); |
340 | | // pending data policy |
341 | | DEFINE_mInt32(pending_data_expire_time_sec, "1800"); |
342 | | // inc_rowset snapshot rs sweep time interval |
343 | | DEFINE_mInt32(tablet_rowset_stale_sweep_time_sec, "600"); |
344 | | // tablet stale rowset sweep by threshold size |
345 | | DEFINE_Bool(tablet_rowset_stale_sweep_by_size, "false"); |
346 | | DEFINE_mInt32(tablet_rowset_stale_sweep_threshold_size, "100"); |
347 | | // garbage sweep policy |
348 | | DEFINE_Int32(max_garbage_sweep_interval, "3600"); |
349 | | DEFINE_Int32(min_garbage_sweep_interval, "180"); |
350 | | DEFINE_mInt32(garbage_sweep_batch_size, "100"); |
351 | | DEFINE_mInt32(snapshot_expire_time_sec, "172800"); |
352 | | // It is only a recommended value. When the disk space is insufficient, |
353 | | // the file storage period under trash dose not have to comply with this parameter. |
354 | | DEFINE_mInt32(trash_file_expire_time_sec, "0"); |
355 | | // minimum file descriptor number |
356 | | // modify them upon necessity |
357 | | DEFINE_Int32(min_file_descriptor_number, "60000"); |
358 | | DEFINE_mBool(disable_segment_cache, "false"); |
359 | | DEFINE_String(row_cache_mem_limit, "20%"); |
360 | | |
361 | | // Cache for storage page size |
362 | | DEFINE_String(storage_page_cache_limit, "20%"); |
363 | | // Shard size for page cache, the value must be power of two. |
364 | | // It's recommended to set it to a value close to the number of BE cores in order to reduce lock contentions. |
365 | | DEFINE_Int32(storage_page_cache_shard_size, "256"); |
366 | | // Percentage for index page cache |
367 | | // all storage page cache will be divided into data_page_cache and index_page_cache |
368 | | DEFINE_Int32(index_page_cache_percentage, "10"); |
369 | | // whether to disable page cache feature in storage |
370 | | DEFINE_mBool(disable_storage_page_cache, "false"); |
371 | | // whether to disable row cache feature in storage |
372 | | DEFINE_mBool(disable_storage_row_cache, "true"); |
373 | | // whether to disable pk page cache feature in storage |
374 | | DEFINE_Bool(disable_pk_storage_page_cache, "false"); |
375 | | |
376 | | // Cache for mow primary key storage page size |
377 | | DEFINE_String(pk_storage_page_cache_limit, "10%"); |
378 | | // data page size for primary key index |
379 | | DEFINE_Int32(primary_key_data_page_size, "32768"); |
380 | | |
381 | | DEFINE_mInt32(data_page_cache_stale_sweep_time_sec, "300"); |
382 | | DEFINE_mInt32(index_page_cache_stale_sweep_time_sec, "600"); |
383 | | DEFINE_mInt32(pk_index_page_cache_stale_sweep_time_sec, "600"); |
384 | | |
385 | | DEFINE_Bool(enable_low_cardinality_optimize, "true"); |
386 | | DEFINE_Bool(enable_low_cardinality_cache_code, "true"); |
387 | | |
388 | | // be policy |
389 | | // whether check compaction checksum |
390 | | DEFINE_mBool(enable_compaction_checksum, "false"); |
391 | | // whether disable automatic compaction task |
392 | | DEFINE_mBool(disable_auto_compaction, "false"); |
393 | | // whether enable vertical compaction |
394 | | DEFINE_mBool(enable_vertical_compaction, "true"); |
395 | | // whether enable ordered data compaction |
396 | | DEFINE_mBool(enable_ordered_data_compaction, "true"); |
397 | | // In vertical compaction, column number for every group |
398 | | DEFINE_mInt32(vertical_compaction_num_columns_per_group, "5"); |
399 | | // In vertical compaction, max memory usage for row_source_buffer |
400 | | DEFINE_Int32(vertical_compaction_max_row_source_memory_mb, "1024"); |
401 | | // In vertical compaction, max dest segment file size |
402 | | DEFINE_mInt64(vertical_compaction_max_segment_size, "1073741824"); |
403 | | |
404 | | // If enabled, segments will be flushed column by column |
405 | | DEFINE_mBool(enable_vertical_segment_writer, "true"); |
406 | | |
407 | | // In ordered data compaction, min segment size for input rowset |
408 | | DEFINE_mInt32(ordered_data_compaction_min_segment_size, "10485760"); |
409 | | |
410 | | // This config can be set to limit thread number in compaction thread pool. |
411 | | DEFINE_mInt32(max_base_compaction_threads, "4"); |
412 | | DEFINE_mInt32(max_cumu_compaction_threads, "-1"); |
413 | | DEFINE_mInt32(max_single_replica_compaction_threads, "-1"); |
414 | | |
415 | | DEFINE_Bool(enable_base_compaction_idle_sched, "true"); |
416 | | DEFINE_mInt64(base_compaction_min_rowset_num, "5"); |
417 | | DEFINE_mInt64(base_compaction_max_compaction_score, "20"); |
418 | | DEFINE_mDouble(base_compaction_min_data_ratio, "0.3"); |
419 | | DEFINE_mInt64(base_compaction_dup_key_max_file_size_mbytes, "1024"); |
420 | | |
421 | | DEFINE_Bool(enable_skip_tablet_compaction, "false"); |
422 | | // output rowset of cumulative compaction total disk size exceed this config size, |
423 | | // this rowset will be given to base compaction, unit is m byte. |
424 | | DEFINE_mInt64(compaction_promotion_size_mbytes, "1024"); |
425 | | |
426 | | // output rowset of cumulative compaction total disk size exceed this config ratio of |
427 | | // base rowset's total disk size, this rowset will be given to base compaction. The value must be between |
428 | | // 0 and 1. |
429 | | DEFINE_mDouble(compaction_promotion_ratio, "0.05"); |
430 | | |
431 | | // the smallest size of rowset promotion. When the rowset is less than this config, this |
432 | | // rowset will be not given to base compaction. The unit is m byte. |
433 | | DEFINE_mInt64(compaction_promotion_min_size_mbytes, "128"); |
434 | | |
435 | | // When output rowset of cumulative compaction total version count (end_version - start_version) |
436 | | // exceed this config count, the rowset will be moved to base compaction |
437 | | // NOTE: this config will work for unique key merge-on-write table only, to reduce version count |
438 | | // related cost on delete bitmap more effectively. |
439 | | DEFINE_mInt64(compaction_promotion_version_count, "1000"); |
440 | | |
441 | | // The lower bound size to do cumulative compaction. When total disk size of candidate rowsets is less than |
442 | | // this size, size_based policy may not do to cumulative compaction. The unit is m byte. |
443 | | DEFINE_mInt64(compaction_min_size_mbytes, "64"); |
444 | | |
445 | | // cumulative compaction policy: min and max delta file's number |
446 | | DEFINE_mInt64(cumulative_compaction_min_deltas, "5"); |
447 | | DEFINE_mInt64(cumulative_compaction_max_deltas, "1000"); |
448 | | DEFINE_mInt32(cumulative_compaction_max_deltas_factor, "10"); |
449 | | |
450 | | // This config can be set to limit thread number in multiget thread pool. |
451 | | DEFINE_mInt32(multi_get_max_threads, "10"); |
452 | | |
453 | | // The upper limit of "permits" held by all compaction tasks. This config can be set to limit memory consumption for compaction. |
454 | | DEFINE_mInt64(total_permits_for_compaction_score, "10000"); |
455 | | |
456 | | // sleep interval in ms after generated compaction tasks |
457 | | DEFINE_mInt32(generate_compaction_tasks_interval_ms, "10"); |
458 | | |
459 | | // sleep interval in second after update replica infos |
460 | | DEFINE_mInt32(update_replica_infos_interval_seconds, "60"); |
461 | | |
462 | | // Compaction task number per disk. |
463 | | // Must be greater than 2, because Base compaction and Cumulative compaction have at least one thread each. |
464 | | DEFINE_mInt32(compaction_task_num_per_disk, "4"); |
465 | | // compaction thread num for fast disk(typically .SSD), must be greater than 2. |
466 | | DEFINE_mInt32(compaction_task_num_per_fast_disk, "8"); |
467 | | DEFINE_Validator(compaction_task_num_per_disk, |
468 | | [](const int config) -> bool { return config >= 2; }); |
469 | | DEFINE_Validator(compaction_task_num_per_fast_disk, |
470 | | [](const int config) -> bool { return config >= 2; }); |
471 | | DEFINE_Validator(low_priority_compaction_task_num_per_disk, |
472 | | [](const int config) -> bool { return config >= 2; }); |
473 | | |
474 | | // How many rounds of cumulative compaction for each round of base compaction when compaction tasks generation. |
475 | | DEFINE_mInt32(cumulative_compaction_rounds_for_each_base_compaction_round, "9"); |
476 | | |
477 | | // Not compact the invisible versions, but with some limitations: |
478 | | // if not timeout, keep no more than compaction_keep_invisible_version_max_count versions; |
479 | | // if timeout, keep no more than compaction_keep_invisible_version_min_count versions. |
480 | | DEFINE_mInt32(compaction_keep_invisible_version_timeout_sec, "1800"); |
481 | | DEFINE_mInt32(compaction_keep_invisible_version_min_count, "50"); |
482 | | DEFINE_mInt32(compaction_keep_invisible_version_max_count, "500"); |
483 | | |
484 | | // Threshold to logging compaction trace, in seconds. |
485 | | DEFINE_mInt32(base_compaction_trace_threshold, "60"); |
486 | | DEFINE_mInt32(cumulative_compaction_trace_threshold, "10"); |
487 | | DEFINE_mBool(disable_compaction_trace_log, "true"); |
488 | | |
489 | | // Interval to picking rowset to compact, in seconds |
490 | | DEFINE_mInt64(pick_rowset_to_compact_interval_sec, "86400"); |
491 | | |
492 | | // Compaction priority schedule |
493 | | DEFINE_mBool(enable_compaction_priority_scheduling, "true"); |
494 | | DEFINE_mInt32(low_priority_compaction_task_num_per_disk, "2"); |
495 | | DEFINE_mInt32(low_priority_compaction_score_threshold, "200"); |
496 | | |
497 | | // Thread count to do tablet meta checkpoint, -1 means use the data directories count. |
498 | | DEFINE_Int32(max_meta_checkpoint_threads, "-1"); |
499 | | |
500 | | // Threshold to logging agent task trace, in seconds. |
501 | | DEFINE_mInt32(agent_task_trace_threshold_sec, "2"); |
502 | | |
503 | | // This config can be set to limit thread number in tablet migration thread pool. |
504 | | DEFINE_Int32(min_tablet_migration_threads, "1"); |
505 | | DEFINE_Int32(max_tablet_migration_threads, "1"); |
506 | | |
507 | | DEFINE_mInt32(finished_migration_tasks_size, "10000"); |
508 | | // If size less than this, the remaining rowsets will be force to complete |
509 | | DEFINE_mInt32(migration_remaining_size_threshold_mb, "10"); |
510 | | // If the task runs longer than this time, the task will be terminated, in seconds. |
511 | | // timeout = std::max(migration_task_timeout_secs, tablet size / 1MB/s) |
512 | | DEFINE_mInt32(migration_task_timeout_secs, "300"); |
513 | | // timeout for try_lock migration lock |
514 | | DEFINE_Int64(migration_lock_timeout_ms, "1000"); |
515 | | |
516 | | // Port to start debug webserver on |
517 | | DEFINE_Int32(webserver_port, "8040"); |
518 | | // Https enable flag |
519 | | DEFINE_Bool(enable_https, "false"); |
520 | | // Path of certificate |
521 | | DEFINE_String(ssl_certificate_path, ""); |
522 | | // Path of private key |
523 | | DEFINE_String(ssl_private_key_path, ""); |
524 | | // Whether to check authorization |
525 | | DEFINE_Bool(enable_all_http_auth, "false"); |
526 | | // Number of webserver workers |
527 | | DEFINE_Int32(webserver_num_workers, "48"); |
528 | | |
529 | | DEFINE_Bool(enable_single_replica_load, "true"); |
530 | | // Number of download workers for single replica load |
531 | | DEFINE_Int32(single_replica_load_download_num_workers, "64"); |
532 | | |
533 | | // Used for mini Load. mini load data file will be removed after this time. |
534 | | DEFINE_Int64(load_data_reserve_hours, "4"); |
535 | | // log error log will be removed after this time |
536 | | DEFINE_mInt64(load_error_log_reserve_hours, "48"); |
537 | | // error log size limit, default 200MB |
538 | | DEFINE_mInt64(load_error_log_limit_bytes, "209715200"); |
539 | | |
540 | | DEFINE_Int32(brpc_heavy_work_pool_threads, "-1"); |
541 | | DEFINE_Int32(brpc_light_work_pool_threads, "-1"); |
542 | | DEFINE_Int32(brpc_heavy_work_pool_max_queue_size, "-1"); |
543 | | DEFINE_Int32(brpc_light_work_pool_max_queue_size, "-1"); |
544 | | DEFINE_mBool(enable_bthread_transmit_block, "true"); |
545 | | DEFINE_Int32(brpc_arrow_flight_work_pool_threads, "-1"); |
546 | | DEFINE_Int32(brpc_arrow_flight_work_pool_max_queue_size, "-1"); |
547 | | |
548 | | //Enable brpc builtin services, see: |
549 | | //https://brpc.apache.org/docs/server/basics/#disable-built-in-services-completely |
550 | | DEFINE_Bool(enable_brpc_builtin_services, "true"); |
551 | | |
552 | | // Enable brpc connection check |
553 | | DEFINE_Bool(enable_brpc_connection_check, "false"); |
554 | | |
555 | | // The maximum amount of data that can be processed by a stream load |
556 | | DEFINE_mInt64(streaming_load_max_mb, "102400"); |
557 | | // Some data formats, such as JSON, cannot be streamed. |
558 | | // Therefore, it is necessary to limit the maximum number of |
559 | | // such data when using stream load to prevent excessive memory consumption. |
560 | | DEFINE_mInt64(streaming_load_json_max_mb, "100"); |
561 | | // the alive time of a TabletsChannel. |
562 | | // If the channel does not receive any data till this time, |
563 | | // the channel will be removed. |
564 | | DEFINE_mInt32(streaming_load_rpc_max_alive_time_sec, "1200"); |
565 | | // the timeout of a rpc to open the tablet writer in remote BE. |
566 | | // short operation time, can set a short timeout |
567 | | DEFINE_Int32(tablet_writer_open_rpc_timeout_sec, "60"); |
568 | | // You can ignore brpc error '[E1011]The server is overcrowded' when writing data. |
569 | | DEFINE_mBool(tablet_writer_ignore_eovercrowded, "true"); |
570 | | DEFINE_mInt32(slave_replica_writer_rpc_timeout_sec, "60"); |
571 | | // Whether to enable stream load record function, the default is false. |
572 | | // False: disable stream load record |
573 | | DEFINE_mBool(enable_stream_load_record, "false"); |
574 | | // batch size of stream load record reported to FE |
575 | | DEFINE_mInt32(stream_load_record_batch_size, "50"); |
576 | | // expire time of stream load record in rocksdb. |
577 | | DEFINE_Int32(stream_load_record_expire_time_secs, "28800"); |
578 | | // time interval to clean expired stream load records |
579 | | DEFINE_mInt64(clean_stream_load_record_interval_secs, "1800"); |
580 | | // enable stream load commit txn on BE directly, bypassing FE. Only for cloud. |
581 | | DEFINE_mBool(enable_stream_load_commit_txn_on_be, "false"); |
582 | | // The buffer size to store stream table function schema info |
583 | | DEFINE_Int64(stream_tvf_buffer_size, "1048576"); // 1MB |
584 | | |
585 | | // OlapTableSink sender's send interval, should be less than the real response time of a tablet writer rpc. |
586 | | // You may need to lower the speed when the sink receiver bes are too busy. |
587 | | DEFINE_mInt32(olap_table_sink_send_interval_microseconds, "1000"); |
588 | | DEFINE_mDouble(olap_table_sink_send_interval_auto_partition_factor, "0.001"); |
589 | | |
590 | | // Fragment thread pool |
591 | | DEFINE_Int32(fragment_mgr_asynic_work_pool_thread_num_min, "16"); |
592 | | DEFINE_Int32(fragment_mgr_asynic_work_pool_thread_num_max, "512"); |
593 | | DEFINE_Int32(fragment_mgr_asynic_work_pool_queue_size, "4096"); |
594 | | |
595 | | // Control the number of disks on the machine. If 0, this comes from the system settings. |
596 | | DEFINE_Int32(num_disks, "0"); |
597 | | // The read size is the size of the reads sent to os. |
598 | | // There is a trade off of latency and throughout, trying to keep disks busy but |
599 | | // not introduce seeks. The literature seems to agree that with 8 MB reads, random |
600 | | // io and sequential io perform similarly. |
601 | | DEFINE_Int32(read_size, "8388608"); // 8 * 1024 * 1024, Read Size (in bytes) |
602 | | DEFINE_Int32(min_buffer_size, "1024"); // 1024, The minimum read buffer size (in bytes) |
603 | | |
604 | | // for pprof |
605 | | DEFINE_String(pprof_profile_dir, "${DORIS_HOME}/log"); |
606 | | // for jeprofile in jemalloc |
607 | | DEFINE_mString(jeprofile_dir, "${DORIS_HOME}/log"); |
608 | | DEFINE_mBool(enable_je_purge_dirty_pages, "true"); |
609 | | DEFINE_mString(je_dirty_pages_mem_limit_percent, "2%"); |
610 | | |
611 | | // to forward compatibility, will be removed later |
612 | | DEFINE_mBool(enable_token_check, "true"); |
613 | | |
614 | | // to open/close system metrics |
615 | | DEFINE_Bool(enable_system_metrics, "true"); |
616 | | |
617 | | // Number of cores Doris will used, this will effect only when it's greater than 0. |
618 | | // Otherwise, Doris will use all cores returned from "/proc/cpuinfo". |
619 | | DEFINE_Int32(num_cores, "0"); |
620 | | |
621 | | // When BE start, If there is a broken disk, BE process will exit by default. |
622 | | // Otherwise, we will ignore the broken disk, |
623 | | DEFINE_Bool(ignore_broken_disk, "false"); |
624 | | |
625 | | // Sleep time in milliseconds between memory maintenance iterations |
626 | | DEFINE_mInt32(memory_maintenance_sleep_time_ms, "20"); |
627 | | |
628 | | // After full gc, no longer full gc and minor gc during sleep. |
629 | | // After minor gc, no minor gc during sleep, but full gc is possible. |
630 | | DEFINE_mInt32(memory_gc_sleep_time_ms, "500"); |
631 | | |
632 | | // max write buffer size before flush, default 200MB |
633 | | DEFINE_mInt64(write_buffer_size, "209715200"); |
634 | | // max buffer size used in memtable for the aggregated table, default 400MB |
635 | | DEFINE_mInt64(write_buffer_size_for_agg, "419430400"); |
636 | | // max parallel flush task per memtable writer |
637 | | DEFINE_mInt32(memtable_flush_running_count_limit, "2"); |
638 | | |
639 | | DEFINE_Int32(load_process_max_memory_limit_percent, "50"); // 50% |
640 | | |
641 | | // If the memory consumption of load jobs exceed load_process_max_memory_limit, |
642 | | // all load jobs will hang there to wait for memtable flush. We should have a |
643 | | // soft limit which can trigger the memtable flush for the load channel who |
644 | | // consumes lagest memory size before we reach the hard limit. The soft limit |
645 | | // might avoid all load jobs hang at the same time. |
646 | | DEFINE_Int32(load_process_soft_mem_limit_percent, "80"); |
647 | | |
648 | | // If load memory consumption is within load_process_safe_mem_permit_percent, |
649 | | // memtable memory limiter will do nothing. |
650 | | DEFINE_Int32(load_process_safe_mem_permit_percent, "5"); |
651 | | |
652 | | // result buffer cancelled time (unit: second) |
653 | | DEFINE_mInt32(result_buffer_cancelled_interval_time, "300"); |
654 | | |
655 | | // arrow flight result sink buffer rows size, default 4096 * 8 |
656 | | DEFINE_mInt32(arrow_flight_result_sink_buffer_size_rows, "32768"); |
657 | | // The timeout for ADBC Client to wait for data using arrow flight reader. |
658 | | // If the query is very complex and no result is generated after this time, consider increasing this timeout. |
659 | | DEFINE_mInt32(arrow_flight_reader_brpc_controller_timeout_ms, "300000"); |
660 | | |
661 | | // the increased frequency of priority for remaining tasks in BlockingPriorityQueue |
662 | | DEFINE_mInt32(priority_queue_remaining_tasks_increased_frequency, "512"); |
663 | | |
664 | | // sync tablet_meta when modifying meta |
665 | | DEFINE_mBool(sync_tablet_meta, "false"); |
666 | | |
667 | | // sync when closing a file writer |
668 | | DEFINE_mBool(sync_file_on_close, "true"); |
669 | | |
670 | | // default thrift rpc timeout ms |
671 | | DEFINE_mInt32(thrift_rpc_timeout_ms, "60000"); |
672 | | |
673 | | // txn commit rpc timeout |
674 | | DEFINE_mInt32(txn_commit_rpc_timeout_ms, "60000"); |
675 | | |
676 | | // If set to true, metric calculator will run |
677 | | DEFINE_Bool(enable_metric_calculator, "true"); |
678 | | |
679 | | // max consumer num in one data consumer group, for routine load |
680 | | DEFINE_mInt32(max_consumer_num_per_group, "3"); |
681 | | |
682 | | // the max size of thread pool for routine load task. |
683 | | // this should be larger than FE config 'max_routine_load_task_num_per_be' (default 5) |
684 | | DEFINE_Int32(max_routine_load_thread_pool_size, "1024"); |
685 | | |
686 | | // max external scan cache batch count, means cache max_memory_cache_batch_count * batch_size row |
687 | | // default is 20, batch_size's default value is 1024 means 20 * 1024 rows will be cached |
688 | | DEFINE_mInt32(max_memory_sink_batch_count, "20"); |
689 | | |
690 | | // This configuration is used for the context gc thread schedule period |
691 | | // note: unit is minute, default is 5min |
692 | | DEFINE_mInt32(scan_context_gc_interval_min, "5"); |
693 | | |
694 | | // es scroll keep-alive |
695 | | DEFINE_String(es_scroll_keepalive, "5m"); |
696 | | |
697 | | // HTTP connection timeout for es |
698 | | DEFINE_mInt32(es_http_timeout_ms, "5000"); |
699 | | |
700 | | // the max client cache number per each host |
701 | | // There are variety of client cache in BE, but currently we use the |
702 | | // same cache size configuration. |
703 | | // TODO(cmy): use different config to set different client cache if necessary. |
704 | | DEFINE_Int32(max_client_cache_size_per_host, "10"); |
705 | | |
706 | | // Dir to save files downloaded by SmallFileMgr |
707 | | DEFINE_String(small_file_dir, "${DORIS_HOME}/lib/small_file/"); |
708 | | // path gc |
709 | | DEFINE_Bool(path_gc_check, "true"); |
710 | | DEFINE_mInt32(path_gc_check_interval_second, "86400"); |
711 | | DEFINE_mInt32(path_gc_check_step, "1000"); |
712 | | DEFINE_mInt32(path_gc_check_step_interval_ms, "10"); |
713 | | |
714 | | // The following 2 configs limit the max usage of disk capacity of a data dir. |
715 | | // If both of these 2 threshold reached, no more data can be writen into that data dir. |
716 | | // The percent of max used capacity of a data dir |
717 | | DEFINE_mInt32(storage_flood_stage_usage_percent, "90"); // 90% |
718 | | // The min bytes that should be left of a data dir |
719 | | DEFINE_mInt64(storage_flood_stage_left_capacity_bytes, "1073741824"); // 1GB |
720 | | // number of thread for flushing memtable per store |
721 | | DEFINE_Int32(flush_thread_num_per_store, "6"); |
722 | | // number of thread for flushing memtable per store, for high priority load task |
723 | | DEFINE_Int32(high_priority_flush_thread_num_per_store, "6"); |
724 | | // number of threads = min(flush_thread_num_per_store * num_store, |
725 | | // max_flush_thread_num_per_cpu * num_cpu) |
726 | | DEFINE_Int32(max_flush_thread_num_per_cpu, "4"); |
727 | | |
728 | | DEFINE_mInt32(wg_flush_thread_num_per_store, "6"); |
729 | | DEFINE_mInt32(wg_flush_thread_num_per_cpu, "4"); |
730 | | |
731 | | // config for tablet meta checkpoint |
732 | | DEFINE_mInt32(tablet_meta_checkpoint_min_new_rowsets_num, "10"); |
733 | | DEFINE_mInt32(tablet_meta_checkpoint_min_interval_secs, "600"); |
734 | | DEFINE_Int32(generate_tablet_meta_checkpoint_tasks_interval_secs, "600"); |
735 | | |
736 | | // config for default rowset type |
737 | | // Valid configs: ALPHA, BETA |
738 | | DEFINE_String(default_rowset_type, "BETA"); |
739 | | |
740 | | // Maximum size of a single message body in all protocols |
741 | | DEFINE_Int64(brpc_max_body_size, "3147483648"); |
742 | | DEFINE_Int64(brpc_socket_max_unwritten_bytes, "-1"); |
743 | | // TODO(zxy): expect to be true in v1.3 |
744 | | // Whether to embed the ProtoBuf Request serialized string together with Tuple/Block data into |
745 | | // Controller Attachment and send it through http brpc when the length of the Tuple/Block data |
746 | | // is greater than 1.8G. This is to avoid the error of Request length overflow (2G). |
747 | | DEFINE_mBool(transfer_large_data_by_brpc, "true"); |
748 | | |
749 | | // max number of txns for every txn_partition_map in txn manager |
750 | | // this is a self protection to avoid too many txns saving in manager |
751 | | DEFINE_mInt64(max_runnings_transactions_per_txn_map, "2000"); |
752 | | |
753 | | // tablet_map_lock shard size, the value is 2^n, n=0,1,2,3,4 |
754 | | // this is a an enhancement for better performance to manage tablet |
755 | | DEFINE_Int32(tablet_map_shard_size, "256"); |
756 | | |
757 | | // txn_map_lock shard size, the value is 2^n, n=0,1,2,3,4 |
758 | | // this is a an enhancement for better performance to manage txn |
759 | | DEFINE_Int32(txn_map_shard_size, "1024"); |
760 | | |
761 | | // txn_lock shard size, the value is 2^n, n=0,1,2,3,4 |
762 | | // this is a an enhancement for better performance to commit and publish txn |
763 | | DEFINE_Int32(txn_shard_size, "1024"); |
764 | | |
765 | | // Whether to continue to start be when load tablet from header failed. |
766 | | DEFINE_Bool(ignore_load_tablet_failure, "false"); |
767 | | |
768 | | // Whether to continue to start be when load tablet from header failed. |
769 | | DEFINE_mBool(ignore_rowset_stale_unconsistent_delete, "false"); |
770 | | |
771 | | // Set max cache's size of query results, the unit is M byte |
772 | | DEFINE_Int32(query_cache_max_size_mb, "256"); |
773 | | |
774 | | // Cache memory is pruned when reach query_cache_max_size_mb + query_cache_elasticity_size_mb |
775 | | DEFINE_Int32(query_cache_elasticity_size_mb, "128"); |
776 | | |
777 | | // Maximum number of cache partitions corresponding to a SQL |
778 | | DEFINE_Int32(query_cache_max_partition_count, "1024"); |
779 | | |
780 | | // Maximum number of version of a tablet. If the version num of a tablet exceed limit, |
781 | | // the load process will reject new incoming load job of this tablet. |
782 | | // This is to avoid too many version num. |
783 | | DEFINE_mInt32(max_tablet_version_num, "2000"); |
784 | | |
785 | | // Frontend mainly use two thrift sever type: THREAD_POOL, THREADED_SELECTOR. if fe use THREADED_SELECTOR model for thrift server, |
786 | | // the thrift_server_type_of_fe should be set THREADED_SELECTOR to make be thrift client to fe constructed with TFramedTransport |
787 | | DEFINE_String(thrift_server_type_of_fe, "THREAD_POOL"); |
788 | | |
789 | | // disable zone map index when page row is too few |
790 | | DEFINE_mInt32(zone_map_row_num_threshold, "20"); |
791 | | |
792 | | // aws sdk log level |
793 | | // Off = 0, |
794 | | // Fatal = 1, |
795 | | // Error = 2, |
796 | | // Warn = 3, |
797 | | // Info = 4, |
798 | | // Debug = 5, |
799 | | // Trace = 6 |
800 | | // Default to turn off aws sdk log, because aws sdk errors that need to be cared will be output through Doris logs |
801 | | DEFINE_Int32(aws_log_level, "0"); |
802 | | |
803 | | // the buffer size when read data from remote storage like s3 |
804 | | DEFINE_mInt32(remote_storage_read_buffer_mb, "16"); |
805 | | |
806 | | // The minimum length when TCMalloc Hook consumes/releases MemTracker, consume size |
807 | | // smaller than this value will continue to accumulate. specified as number of bytes. |
808 | | // Decreasing this value will increase the frequency of consume/release. |
809 | | // Increasing this value will cause MemTracker statistics to be inaccurate. |
810 | | DEFINE_mInt32(mem_tracker_consume_min_size_bytes, "1048576"); |
811 | | |
812 | | // The version information of the tablet will be stored in the memory |
813 | | // in an adjacency graph data structure. |
814 | | // And as the new version is written and the old version is deleted, |
815 | | // the data structure will begin to have empty vertex with no edge associations(orphan vertex). |
816 | | // This config is used to control that when the proportion of orphan vertex is greater than the threshold, |
817 | | // the adjacency graph will be rebuilt to ensure that the data structure will not expand indefinitely. |
818 | | // This config usually only needs to be modified during testing. |
819 | | // In most cases, it does not need to be modified. |
820 | | DEFINE_mDouble(tablet_version_graph_orphan_vertex_ratio, "0.1"); |
821 | | |
822 | | // share delta writers when memtable_on_sink_node = true |
823 | | DEFINE_Bool(share_delta_writers, "true"); |
824 | | // timeout for open load stream rpc in ms |
825 | | DEFINE_Int64(open_load_stream_timeout_ms, "60000"); // 60s |
826 | | // enable write background when using brpc stream |
827 | | DEFINE_mBool(enable_brpc_stream_write_background, "true"); |
828 | | |
829 | | // brpc streaming max_buf_size in bytes |
830 | | DEFINE_Int64(load_stream_max_buf_size, "20971520"); // 20MB |
831 | | // brpc streaming messages_in_batch |
832 | | DEFINE_Int32(load_stream_messages_in_batch, "128"); |
833 | | // brpc streaming StreamWait seconds on EAGAIN |
834 | | DEFINE_Int32(load_stream_eagain_wait_seconds, "600"); |
835 | | // max tasks per flush token in load stream |
836 | | DEFINE_Int32(load_stream_flush_token_max_tasks, "15"); |
837 | | // max wait flush token time in load stream |
838 | | DEFINE_Int32(load_stream_max_wait_flush_token_time_ms, "600000"); |
839 | | // number of send batch thread pool size |
840 | | DEFINE_Int32(send_batch_thread_pool_thread_num, "64"); |
841 | | // number of send batch thread pool queue size |
842 | | DEFINE_Int32(send_batch_thread_pool_queue_size, "102400"); |
843 | | |
844 | | // Limit the number of segment of a newly created rowset. |
845 | | // The newly created rowset may to be compacted after loading, |
846 | | // so if there are too many segment in a rowset, the compaction process |
847 | | // will run out of memory. |
848 | | // When doing compaction, each segment may take at least 1MB buffer. |
849 | | DEFINE_mInt32(max_segment_num_per_rowset, "1000"); |
850 | | DEFINE_mInt32(segment_compression_threshold_kb, "256"); |
851 | | |
852 | | // The connection timeout when connecting to external table such as odbc table. |
853 | | DEFINE_mInt32(external_table_connect_timeout_sec, "30"); |
854 | | |
855 | | // Time to clean up useless JDBC connection pool cache |
856 | | DEFINE_mInt32(jdbc_connection_pool_cache_clear_time_sec, "28800"); |
857 | | |
858 | | // Global bitmap cache capacity for aggregation cache, size in bytes |
859 | | DEFINE_Int64(delete_bitmap_agg_cache_capacity, "104857600"); |
860 | | // The default delete bitmap cache is set to 100MB, |
861 | | // which can be insufficient and cause performance issues when the amount of user data is large. |
862 | | // To mitigate the problem of an inadequate cache, |
863 | | // we will take the larger of 0.5% of the total memory and 100MB as the delete bitmap cache size. |
864 | | DEFINE_String(delete_bitmap_dynamic_agg_cache_limit, "0.5%"); |
865 | | DEFINE_mInt32(delete_bitmap_agg_cache_stale_sweep_time_sec, "1800"); |
866 | | |
867 | | // reference https://github.com/edenhill/librdkafka/blob/master/INTRODUCTION.md#broker-version-compatibility |
868 | | // If the dependent kafka broker version older than 0.10.0.0, |
869 | | // the value of kafka_api_version_request should be false, and the |
870 | | // value set by the fallback version kafka_broker_version_fallback will be used, |
871 | | // and the valid values are: 0.9.0.x, 0.8.x.y. |
872 | | DEFINE_String(kafka_api_version_request, "true"); |
873 | | DEFINE_String(kafka_broker_version_fallback, "0.10.0"); |
874 | | DEFINE_String(kafka_debug, "disable"); |
875 | | |
876 | | // The number of pool siz of routine load consumer. |
877 | | // If you meet the error describe in https://github.com/edenhill/librdkafka/issues/3608 |
878 | | // Change this size to 0 to fix it temporarily. |
879 | | DEFINE_mInt32(routine_load_consumer_pool_size, "1024"); |
880 | | |
881 | | // Used in single-stream-multi-table load. When receive a batch of messages from kafka, |
882 | | // if the size of batch is more than this threshold, we will request plans for all related tables. |
883 | | DEFINE_Int32(multi_table_batch_plan_threshold, "200"); |
884 | | |
885 | | // Used in single-stream-multi-table load. When receiving a batch of messages from Kafka, |
886 | | // if the size of the table wait for plan is more than this threshold, we will request plans for all related tables. |
887 | | // The param is aimed to avoid requesting and executing too many plans at once. |
888 | | // Performing small batch processing on multiple tables during the loaded process can reduce the pressure of a single RPC |
889 | | // and improve the real-time processing of data. |
890 | | DEFINE_Int32(multi_table_max_wait_tables, "5"); |
891 | | |
892 | | // When the timeout of a load task is less than this threshold, |
893 | | // Doris treats it as a high priority task. |
894 | | // high priority tasks use a separate thread pool for flush and do not block rpc by memory cleanup logic. |
895 | | // this threshold is mainly used to identify routine load tasks and should not be modified if not necessary. |
896 | | DEFINE_mInt32(load_task_high_priority_threshold_second, "120"); |
897 | | |
898 | | // The min timeout of load rpc (add batch, close, etc.) |
899 | | // Because a load rpc may be blocked for a while. |
900 | | // Increase this config may avoid rpc timeout. |
901 | | DEFINE_mInt32(min_load_rpc_timeout_ms, "20000"); |
902 | | |
903 | | // use which protocol to access function service, candicate is baidu_std/h2:grpc |
904 | | DEFINE_String(function_service_protocol, "h2:grpc"); |
905 | | |
906 | | // use which load balancer to select server to connect |
907 | | DEFINE_String(rpc_load_balancer, "rr"); |
908 | | |
909 | | // a soft limit of string type length, the hard limit is 2GB - 4, but if too long will cause very low performance, |
910 | | // so we set a soft limit, default is 1MB |
911 | | DEFINE_mInt32(string_type_length_soft_limit_bytes, "1048576"); |
912 | | |
913 | | DEFINE_Validator(string_type_length_soft_limit_bytes, |
914 | | [](const int config) -> bool { return config > 0 && config <= 2147483643; }); |
915 | | |
916 | | // Threshold of reading a small file into memory |
917 | | DEFINE_mInt32(in_memory_file_size, "1048576"); // 1MB |
918 | | |
919 | | // Max size of parquet page header in bytes |
920 | | DEFINE_mInt32(parquet_header_max_size_mb, "1"); |
921 | | // Max buffer size for parquet row group |
922 | | DEFINE_mInt32(parquet_rowgroup_max_buffer_mb, "128"); |
923 | | // Max buffer size for parquet chunk column |
924 | | DEFINE_mInt32(parquet_column_max_buffer_mb, "8"); |
925 | | DEFINE_mDouble(max_amplified_read_ratio, "0.8"); |
926 | | DEFINE_mInt32(merged_oss_min_io_size, "1048576"); |
927 | | DEFINE_mInt32(merged_hdfs_min_io_size, "8192"); |
928 | | |
929 | | // OrcReader |
930 | | DEFINE_mInt32(orc_natural_read_size_mb, "8"); |
931 | | DEFINE_mInt64(big_column_size_buffer, "65535"); |
932 | | DEFINE_mInt64(small_column_size_buffer, "100"); |
933 | | |
934 | | // Perform the always_true check at intervals determined by runtime_filter_sampling_frequency |
935 | | DEFINE_mInt32(runtime_filter_sampling_frequency, "64"); |
936 | | DEFINE_mInt32(execution_max_rpc_timeout_sec, "3600"); |
937 | | DEFINE_mBool(execution_ignore_eovercrowded, "true"); |
938 | | // cooldown task configs |
939 | | DEFINE_Int32(cooldown_thread_num, "5"); |
940 | | DEFINE_mInt64(generate_cooldown_task_interval_sec, "20"); |
941 | | DEFINE_mInt32(remove_unused_remote_files_interval_sec, "21600"); // 6h |
942 | | DEFINE_mInt32(confirm_unused_remote_files_interval_sec, "60"); |
943 | | DEFINE_Int32(cold_data_compaction_thread_num, "2"); |
944 | | DEFINE_mInt32(cold_data_compaction_interval_sec, "1800"); |
945 | | |
946 | | DEFINE_String(tmp_file_dir, "tmp"); |
947 | | |
948 | | DEFINE_Int32(min_s3_file_system_thread_num, "16"); |
949 | | DEFINE_Int32(max_s3_file_system_thread_num, "64"); |
950 | | |
951 | | DEFINE_Bool(enable_time_lut, "true"); |
952 | | DEFINE_mBool(enable_simdjson_reader, "true"); |
953 | | |
954 | | DEFINE_mBool(enable_query_like_bloom_filter, "true"); |
955 | | // number of s3 scanner thread pool size |
956 | | DEFINE_Int32(doris_remote_scanner_thread_pool_thread_num, "48"); |
957 | | // number of s3 scanner thread pool queue size |
958 | | DEFINE_Int32(doris_remote_scanner_thread_pool_queue_size, "102400"); |
959 | | DEFINE_mInt64(block_cache_wait_timeout_ms, "1000"); |
960 | | DEFINE_mInt64(cache_lock_long_tail_threshold, "1000"); |
961 | | DEFINE_Int64(file_cache_recycle_keys_size, "1000000"); |
962 | | |
963 | | // limit the queue of pending batches which will be sent by a single nodechannel |
964 | | DEFINE_mInt64(nodechannel_pending_queue_max_bytes, "67108864"); |
965 | | |
966 | | // The batch size for sending data by brpc streaming client |
967 | | DEFINE_mInt64(brpc_streaming_client_batch_bytes, "262144"); |
968 | | |
969 | | // Max waiting time to wait the "plan fragment start" rpc. |
970 | | // If timeout, the fragment will be cancelled. |
971 | | // This parameter is usually only used when the FE loses connection, |
972 | | // and the BE can automatically cancel the relevant fragment after the timeout, |
973 | | // so as to avoid occupying the execution thread for a long time. |
974 | | DEFINE_mInt32(max_fragment_start_wait_time_seconds, "30"); |
975 | | |
976 | | DEFINE_mInt32(fragment_mgr_cancel_worker_interval_seconds, "1"); |
977 | | |
978 | | // Node role tag for backend. Mix role is the default role, and computation role have no |
979 | | // any tablet. |
980 | | DEFINE_String(be_node_role, "mix"); |
981 | | |
982 | | // Hide webserver page for safety. |
983 | | // Hide the be config page for webserver. |
984 | | DEFINE_Bool(hide_webserver_config_page, "false"); |
985 | | |
986 | | DEFINE_Bool(enable_segcompaction, "true"); |
987 | | |
988 | | // Max number of segments allowed in a single segcompaction task. |
989 | | DEFINE_Int32(segcompaction_batch_size, "10"); |
990 | | |
991 | | // Max row count allowed in a single source segment, bigger segments will be skipped. |
992 | | DEFINE_Int32(segcompaction_candidate_max_rows, "1048576"); |
993 | | |
994 | | // Max file size allowed in a single source segment, bigger segments will be skipped. |
995 | | DEFINE_Int64(segcompaction_candidate_max_bytes, "104857600"); |
996 | | |
997 | | // Max total row count allowed in a single segcompaction task. |
998 | | DEFINE_Int32(segcompaction_task_max_rows, "1572864"); |
999 | | |
1000 | | // Max total file size allowed in a single segcompaction task. |
1001 | | DEFINE_Int64(segcompaction_task_max_bytes, "157286400"); |
1002 | | |
1003 | | // Global segcompaction thread pool size. |
1004 | | DEFINE_mInt32(segcompaction_num_threads, "5"); |
1005 | | |
1006 | | // enable java udf and jdbc scannode |
1007 | | DEFINE_Bool(enable_java_support, "true"); |
1008 | | |
1009 | | // Set config randomly to check more issues in github workflow |
1010 | | DEFINE_Bool(enable_fuzzy_mode, "false"); |
1011 | | |
1012 | | DEFINE_Bool(enable_debug_points, "false"); |
1013 | | |
1014 | | DEFINE_Int32(pipeline_executor_size, "0"); |
1015 | | DEFINE_Bool(enable_workload_group_for_scan, "false"); |
1016 | | DEFINE_mInt64(workload_group_scan_task_wait_timeout_ms, "10000"); |
1017 | | |
1018 | | // Whether use schema dict in backend side instead of MetaService side(cloud mode) |
1019 | | DEFINE_mBool(variant_use_cloud_schema_dict, "true"); |
1020 | | DEFINE_mDouble(variant_ratio_of_defaults_as_sparse_column, "1"); |
1021 | | DEFINE_mInt64(variant_threshold_rows_to_estimate_sparse_column, "2048"); |
1022 | | DEFINE_mBool(variant_throw_exeception_on_invalid_json, "false"); |
1023 | | |
1024 | | // block file cache |
1025 | | DEFINE_Bool(enable_file_cache, "false"); |
1026 | | // format: [{"path":"/path/to/file_cache","total_size":21474836480,"query_limit":10737418240}] |
1027 | | // format: [{"path":"/path/to/file_cache","total_size":21474836480,"query_limit":10737418240},{"path":"/path/to/file_cache2","total_size":21474836480,"query_limit":10737418240}] |
1028 | | // format: {"path": "/path/to/file_cache", "total_size":53687091200, "ttl_percent":50, "normal_percent":40, "disposable_percent":5, "index_percent":5} |
1029 | | // format: [{"path": "xxx", "total_size":53687091200, "storage": "memory"}] |
1030 | | // Note1: storage is "disk" by default |
1031 | | // Note2: when the storage is "memory", the path is ignored. So you can set xxx to anything you like |
1032 | | // and doris will just reset the path to "memory" internally. |
1033 | | // In a very wierd case when your storage is disk, and the directory, by accident, is named |
1034 | | // "memory" for some reason, you should write the path as: |
1035 | | // {"path": "memory", "total_size":53687091200, "storage": "disk"} |
1036 | | // or use the default storage value: |
1037 | | // {"path": "memory", "total_size":53687091200} |
1038 | | // Both will use the directory "memory" on the disk instead of the real RAM. |
1039 | | DEFINE_String(file_cache_path, "[{\"path\":\"${DORIS_HOME}/file_cache\"}]"); |
1040 | | DEFINE_Int64(file_cache_each_block_size, "1048576"); // 1MB |
1041 | | |
1042 | | DEFINE_Bool(clear_file_cache, "false"); |
1043 | | DEFINE_Bool(enable_file_cache_query_limit, "false"); |
1044 | | DEFINE_mInt32(file_cache_enter_disk_resource_limit_mode_percent, "88"); |
1045 | | DEFINE_mInt32(file_cache_exit_disk_resource_limit_mode_percent, "80"); |
1046 | | DEFINE_mBool(enable_read_cache_file_directly, "false"); |
1047 | | DEFINE_mBool(file_cache_enable_evict_from_other_queue_by_size, "true"); |
1048 | | DEFINE_mInt64(file_cache_ttl_valid_check_interval_second, "0"); // zero for not checking |
1049 | | // If true, evict the ttl cache using LRU when full. |
1050 | | // Otherwise, only expiration can evict ttl and new data won't add to cache when full. |
1051 | | DEFINE_Bool(enable_ttl_cache_evict_using_lru, "true"); |
1052 | | DEFINE_mBool(enbale_dump_error_file, "true"); |
1053 | | // limit the max size of error log on disk |
1054 | | DEFINE_mInt64(file_cache_error_log_limit_bytes, "209715200"); // 200MB |
1055 | | |
1056 | | DEFINE_mInt32(index_cache_entry_stay_time_after_lookup_s, "1800"); |
1057 | | DEFINE_mInt32(inverted_index_cache_stale_sweep_time_sec, "600"); |
1058 | | // inverted index searcher cache size |
1059 | | DEFINE_String(inverted_index_searcher_cache_limit, "10%"); |
1060 | | DEFINE_Bool(enable_inverted_index_cache_check_timestamp, "true"); |
1061 | | DEFINE_Int32(inverted_index_fd_number_limit_percent, "20"); // 20% |
1062 | | DEFINE_Int32(inverted_index_query_cache_shards, "256"); |
1063 | | |
1064 | | // inverted index match bitmap cache size |
1065 | | DEFINE_String(inverted_index_query_cache_limit, "10%"); |
1066 | | |
1067 | | // inverted index |
1068 | | DEFINE_mDouble(inverted_index_ram_buffer_size, "512"); |
1069 | | // -1 indicates not working. |
1070 | | // Normally we should not change this, it's useful for testing. |
1071 | | DEFINE_mInt32(inverted_index_max_buffered_docs, "-1"); |
1072 | | // dict path for chinese analyzer |
1073 | | DEFINE_String(inverted_index_dict_path, "${DORIS_HOME}/dict"); |
1074 | | DEFINE_Int32(inverted_index_read_buffer_size, "4096"); |
1075 | | // tree depth for bkd index |
1076 | | DEFINE_Int32(max_depth_in_bkd_tree, "32"); |
1077 | | // index compaction |
1078 | | DEFINE_mBool(inverted_index_compaction_enable, "true"); |
1079 | | // Only for debug, do not use in production |
1080 | | DEFINE_mBool(debug_inverted_index_compaction, "false"); |
1081 | | // index by RAM directory |
1082 | | DEFINE_mBool(inverted_index_ram_dir_enable, "true"); |
1083 | | // use num_broadcast_buffer blocks as buffer to do broadcast |
1084 | | DEFINE_Int32(num_broadcast_buffer, "32"); |
1085 | | |
1086 | | // max depth of expression tree allowed. |
1087 | | DEFINE_Int32(max_depth_of_expr_tree, "600"); |
1088 | | |
1089 | | // Report a tablet as bad when io errors occurs more than this value. |
1090 | | DEFINE_mInt64(max_tablet_io_errors, "-1"); |
1091 | | |
1092 | | // Report a tablet as bad when its path not found |
1093 | | DEFINE_Int32(tablet_path_check_interval_seconds, "-1"); |
1094 | | DEFINE_mInt32(tablet_path_check_batch_size, "1000"); |
1095 | | |
1096 | | // it must be larger than or equal to 5MB |
1097 | | DEFINE_mInt64(s3_write_buffer_size, "5242880"); |
1098 | | // Log interval when doing s3 upload task |
1099 | | DEFINE_mInt32(s3_file_writer_log_interval_second, "60"); |
1100 | | DEFINE_mInt64(file_cache_max_file_reader_cache_size, "1000000"); |
1101 | | DEFINE_mInt64(hdfs_write_batch_buffer_size_mb, "1"); // 1MB |
1102 | | |
1103 | | //disable shrink memory by default |
1104 | | DEFINE_mBool(enable_shrink_memory, "false"); |
1105 | | DEFINE_mInt32(schema_cache_capacity, "1024"); |
1106 | | DEFINE_mInt32(schema_cache_sweep_time_sec, "100"); |
1107 | | |
1108 | | // max number of segment cache, default -1 for backward compatibility fd_number*2/5 |
1109 | | DEFINE_Int32(segment_cache_capacity, "-1"); |
1110 | | DEFINE_Int32(segment_cache_fd_percentage, "20"); |
1111 | | DEFINE_mInt32(estimated_mem_per_column_reader, "512"); |
1112 | | DEFINE_Int32(segment_cache_memory_percentage, "5"); |
1113 | | |
1114 | | // enable feature binlog, default false |
1115 | | DEFINE_Bool(enable_feature_binlog, "false"); |
1116 | | |
1117 | | // enable set in BitmapValue |
1118 | | DEFINE_Bool(enable_set_in_bitmap_value, "true"); |
1119 | | |
1120 | | DEFINE_Int64(max_hdfs_file_handle_cache_num, "1000"); |
1121 | | DEFINE_Int32(max_hdfs_file_handle_cache_time_sec, "3600"); |
1122 | | DEFINE_Int64(max_external_file_meta_cache_num, "1000"); |
1123 | | DEFINE_mInt32(common_obj_lru_cache_stale_sweep_time_sec, "900"); |
1124 | | // Apply delete pred in cumu compaction |
1125 | | DEFINE_mBool(enable_delete_when_cumu_compaction, "false"); |
1126 | | |
1127 | | // max_write_buffer_number for rocksdb |
1128 | | DEFINE_Int32(rocksdb_max_write_buffer_number, "5"); |
1129 | | |
1130 | | DEFINE_mBool(allow_zero_date, "false"); |
1131 | | DEFINE_Bool(allow_invalid_decimalv2_literal, "false"); |
1132 | | DEFINE_mString(kerberos_ccache_path, ""); |
1133 | | DEFINE_mString(kerberos_krb5_conf_path, "/etc/krb5.conf"); |
1134 | | |
1135 | | DEFINE_mString(get_stack_trace_tool, "libunwind"); |
1136 | | DEFINE_mString(dwarf_location_info_mode, "FAST"); |
1137 | | DEFINE_mBool(enable_address_sanitizers_with_stack_trace, "true"); |
1138 | | |
1139 | | // the ratio of _prefetch_size/_batch_size in AutoIncIDBuffer |
1140 | | DEFINE_mInt64(auto_inc_prefetch_size_ratio, "10"); |
1141 | | |
1142 | | // the ratio of _low_level_water_level_mark/_batch_size in AutoIncIDBuffer |
1143 | | DEFINE_mInt64(auto_inc_low_water_level_mark_size_ratio, "3"); |
1144 | | |
1145 | | // number of threads that fetch auto-inc ranges from FE |
1146 | | DEFINE_mInt64(auto_inc_fetch_thread_num, "3"); |
1147 | | // default max to 2048 connections |
1148 | | DEFINE_mInt64(lookup_connection_cache_capacity, "2048"); |
1149 | | |
1150 | | // level of compression when using LZ4_HC, whose defalut value is LZ4HC_CLEVEL_DEFAULT |
1151 | | DEFINE_mInt64(LZ4_HC_compression_level, "9"); |
1152 | | |
1153 | | DEFINE_mBool(enable_merge_on_write_correctness_check, "true"); |
1154 | | // USED FOR DEBUGING |
1155 | | // core directly if the compaction found there's duplicate key on mow table |
1156 | | DEFINE_mBool(enable_mow_compaction_correctness_check_core, "false"); |
1157 | | // rowid conversion correctness check when compaction for mow table |
1158 | | DEFINE_mBool(enable_rowid_conversion_correctness_check, "false"); |
1159 | | // missing rows correctness check when compaction for mow table |
1160 | | DEFINE_mBool(enable_missing_rows_correctness_check, "false"); |
1161 | | // When the number of missing versions is more than this value, do not directly |
1162 | | // retry the publish and handle it through async publish. |
1163 | | DEFINE_mInt32(mow_publish_max_discontinuous_version_num, "20"); |
1164 | | // When the version is not continuous for MOW table in publish phase and the gap between |
1165 | | // current txn's publishing version and the max version of the tablet exceeds this value, |
1166 | | // don't print warning log |
1167 | | DEFINE_mInt32(publish_version_gap_logging_threshold, "200"); |
1168 | | |
1169 | | // The secure path with user files, used in the `local` table function. |
1170 | | DEFINE_mString(user_files_secure_path, "${DORIS_HOME}"); |
1171 | | |
1172 | | DEFINE_Int32(fe_expire_duration_seconds, "60"); |
1173 | | |
1174 | | DEFINE_Int32(grace_shutdown_wait_seconds, "120"); |
1175 | | |
1176 | | DEFINE_Int16(bitmap_serialize_version, "1"); |
1177 | | |
1178 | | // group commit config |
1179 | | DEFINE_String(group_commit_wal_path, ""); |
1180 | | DEFINE_Int32(group_commit_replay_wal_retry_num, "10"); |
1181 | | DEFINE_Int32(group_commit_replay_wal_retry_interval_seconds, "5"); |
1182 | | DEFINE_Int32(group_commit_replay_wal_retry_interval_max_seconds, "1800"); |
1183 | | DEFINE_Int32(group_commit_relay_wal_threads, "10"); |
1184 | | // This config can be set to limit thread number in group commit request fragment thread pool. |
1185 | | DEFINE_Int32(group_commit_insert_threads, "10"); |
1186 | | DEFINE_Int32(group_commit_memory_rows_for_max_filter_ratio, "10000"); |
1187 | | DEFINE_Bool(wait_internal_group_commit_finish, "false"); |
1188 | | // Max size(bytes) of group commit queues, used for mem back pressure, defult 64M. |
1189 | | DEFINE_mInt32(group_commit_queue_mem_limit, "67108864"); |
1190 | | // Max size(bytes) or percentage(%) of wal disk usage, used for disk space back pressure, default 10% of the disk available space. |
1191 | | // group_commit_wal_max_disk_limit=1024 or group_commit_wal_max_disk_limit=10% can be automatically identified. |
1192 | | DEFINE_String(group_commit_wal_max_disk_limit, "10%"); |
1193 | | DEFINE_Bool(group_commit_wait_replay_wal_finish, "false"); |
1194 | | |
1195 | | DEFINE_mInt32(scan_thread_nice_value, "0"); |
1196 | | DEFINE_mInt32(tablet_schema_cache_recycle_interval, "3600"); |
1197 | | DEFINE_mInt32(tablet_schema_cache_capacity, "102400"); |
1198 | | |
1199 | | DEFINE_Bool(exit_on_exception, "false"); |
1200 | | // This config controls whether the s3 file writer would flush cache asynchronously |
1201 | | DEFINE_Bool(enable_flush_file_cache_async, "true"); |
1202 | | |
1203 | | // cgroup |
1204 | | DEFINE_mString(doris_cgroup_cpu_path, ""); |
1205 | | |
1206 | | DEFINE_mBool(enable_be_proc_monitor, "false"); |
1207 | | DEFINE_mInt32(be_proc_monitor_interval_ms, "10000"); |
1208 | | |
1209 | | DEFINE_mBool(enable_workload_group_memory_gc, "true"); |
1210 | | |
1211 | | DEFINE_Bool(ignore_always_true_predicate_for_segment, "true"); |
1212 | | |
1213 | | // Ingest binlog work pool size, -1 is disable, 0 is hardware concurrency |
1214 | | DEFINE_Int32(ingest_binlog_work_pool_size, "-1"); |
1215 | | |
1216 | | // Download binlog rate limit, unit is KB/s, 0 means no limit |
1217 | | DEFINE_Int32(download_binlog_rate_limit_kbs, "0"); |
1218 | | |
1219 | | DEFINE_mInt32(buffered_reader_read_timeout_ms, "600000"); |
1220 | | |
1221 | | DEFINE_Bool(enable_snapshot_action, "false"); |
1222 | | |
1223 | | DEFINE_mInt32(variant_max_merged_tablet_schema_size, "2048"); |
1224 | | |
1225 | | DEFINE_mBool(enable_column_type_check, "true"); |
1226 | | // 128 MB |
1227 | | DEFINE_mInt64(local_exchange_buffer_mem_limit, "134217728"); |
1228 | | |
1229 | | // Default 300s, if its value <= 0, then log is disabled |
1230 | | DEFINE_mInt64(enable_debug_log_timeout_secs, "0"); |
1231 | | |
1232 | | // Tolerance for the number of partition id 0 in rowset, default 0 |
1233 | | DEFINE_Int32(ignore_invalid_partition_id_rowset_num, "0"); |
1234 | | |
1235 | | DEFINE_mInt32(report_query_statistics_interval_ms, "3000"); |
1236 | | // 30s |
1237 | | DEFINE_mInt32(query_statistics_reserve_timeout_ms, "30000"); |
1238 | | |
1239 | | DEFINE_mInt32(report_exec_status_thread_num, "5"); |
1240 | | |
1241 | | // consider two high usage disk at the same available level if they do not exceed this diff. |
1242 | | DEFINE_mDouble(high_disk_avail_level_diff_usages, "0.15"); |
1243 | | |
1244 | | // create tablet in partition random robin idx lru size, default 10000 |
1245 | | DEFINE_Int32(partition_disk_index_lru_size, "10000"); |
1246 | | // limit the storage space that query spill files can use |
1247 | | DEFINE_String(spill_storage_root_path, ""); |
1248 | | DEFINE_String(spill_storage_limit, "20%"); // 20% |
1249 | | DEFINE_mInt32(spill_gc_interval_ms, "2000"); // 2s |
1250 | | DEFINE_mInt32(spill_gc_work_time_ms, "2000"); // 2s |
1251 | | DEFINE_Int32(spill_io_thread_pool_thread_num, "-1"); |
1252 | | DEFINE_Validator(spill_io_thread_pool_thread_num, [](const int config) -> bool { |
1253 | | if (config == -1) { |
1254 | | CpuInfo::init(); |
1255 | | spill_io_thread_pool_thread_num = std::max(48, CpuInfo::num_cores() * 2); |
1256 | | } |
1257 | | return true; |
1258 | | }); |
1259 | | DEFINE_Int32(spill_io_thread_pool_queue_size, "102400"); |
1260 | | |
1261 | | DEFINE_mBool(check_segment_when_build_rowset_meta, "false"); |
1262 | | |
1263 | | DEFINE_mInt32(max_s3_client_retry, "10"); |
1264 | | DEFINE_mInt32(s3_read_base_wait_time_ms, "100"); |
1265 | | DEFINE_mInt32(s3_read_max_wait_time_ms, "800"); |
1266 | | |
1267 | | DEFINE_mBool(enable_s3_rate_limiter, "false"); |
1268 | | DEFINE_mInt64(s3_get_bucket_tokens, "1000000000000000000"); |
1269 | | DEFINE_mInt64(s3_get_token_per_second, "1000000000000000000"); |
1270 | | DEFINE_mInt64(s3_get_token_limit, "0"); |
1271 | | |
1272 | | DEFINE_mInt64(s3_put_bucket_tokens, "1000000000000000000"); |
1273 | | DEFINE_mInt64(s3_put_token_per_second, "1000000000000000000"); |
1274 | | DEFINE_mInt64(s3_put_token_limit, "0"); |
1275 | | |
1276 | | DEFINE_String(trino_connector_plugin_dir, "${DORIS_HOME}/connectors"); |
1277 | | |
1278 | | // ca_cert_file is in this path by default, Normally no modification is required |
1279 | | // ca cert default path is different from different OS |
1280 | | DEFINE_mString(ca_cert_file_paths, |
1281 | | "/etc/pki/tls/certs/ca-bundle.crt;/etc/ssl/certs/ca-certificates.crt;" |
1282 | | "/etc/ssl/ca-bundle.pem"); |
1283 | | |
1284 | | /** Table sink configurations(currently contains only external table types) **/ |
1285 | | // Minimum data processed to scale writers in exchange when non partition writing |
1286 | | DEFINE_mInt64(table_sink_non_partition_write_scaling_data_processed_threshold, |
1287 | | "26214400"); // 25MB |
1288 | | // Minimum data processed to trigger skewed partition rebalancing in exchange when partition writing |
1289 | | DEFINE_mInt64(table_sink_partition_write_min_data_processed_rebalance_threshold, |
1290 | | "26214400"); // 25MB |
1291 | | // Minimum partition data processed to rebalance writers in exchange when partition writing |
1292 | | DEFINE_mInt64(table_sink_partition_write_min_partition_data_processed_rebalance_threshold, |
1293 | | "15728640"); // 15MB |
1294 | | // Maximum processed partition nums of per writer when partition writing |
1295 | | DEFINE_mInt32(table_sink_partition_write_max_partition_nums_per_writer, "128"); |
1296 | | |
1297 | | /** Hive sink configurations **/ |
1298 | | DEFINE_mInt64(hive_sink_max_file_size, "1073741824"); // 1GB |
1299 | | |
1300 | | /** Iceberg sink configurations **/ |
1301 | | DEFINE_mInt64(iceberg_sink_max_file_size, "1073741824"); // 1GB |
1302 | | |
1303 | | DEFINE_mInt32(thrift_client_open_num_tries, "1"); |
1304 | | |
1305 | | DEFINE_Bool(enable_index_compaction, "false"); |
1306 | | |
1307 | | // http scheme in S3Client to use. E.g. http or https |
1308 | | DEFINE_String(s3_client_http_scheme, "http"); |
1309 | | DEFINE_Validator(s3_client_http_scheme, [](const std::string& config) -> bool { |
1310 | | return config == "http" || config == "https"; |
1311 | | }); |
1312 | | |
1313 | | DEFINE_mBool(ignore_schema_change_check, "false"); |
1314 | | |
1315 | | DEFINE_mInt64(string_overflow_size, "4294967295"); // std::numic_limits<uint32_t>::max() |
1316 | | |
1317 | | // The min thread num for BufferedReaderPrefetchThreadPool |
1318 | | DEFINE_Int64(num_buffered_reader_prefetch_thread_pool_min_thread, "16"); |
1319 | | // The max thread num for BufferedReaderPrefetchThreadPool |
1320 | | DEFINE_Int64(num_buffered_reader_prefetch_thread_pool_max_thread, "64"); |
1321 | | // The min thread num for S3FileUploadThreadPool |
1322 | | DEFINE_Int64(num_s3_file_upload_thread_pool_min_thread, "16"); |
1323 | | // The max thread num for S3FileUploadThreadPool |
1324 | | DEFINE_Int64(num_s3_file_upload_thread_pool_max_thread, "64"); |
1325 | | // The maximum jvm heap usage ratio for hdfs write workload |
1326 | | DEFINE_mDouble(max_hdfs_wirter_jni_heap_usage_ratio, "0.5"); |
1327 | | // The sleep milliseconds duration when hdfs write exceeds the maximum usage |
1328 | | DEFINE_mInt64(hdfs_jni_write_sleep_milliseconds, "300"); |
1329 | | // The max retry times when hdfs write failed |
1330 | | DEFINE_mInt64(hdfs_jni_write_max_retry_time, "3"); |
1331 | | |
1332 | | // The min thread num for NonBlockCloseThreadPool |
1333 | | DEFINE_Int64(min_nonblock_close_thread_num, "12"); |
1334 | | // The max thread num for NonBlockCloseThreadPool |
1335 | | DEFINE_Int64(max_nonblock_close_thread_num, "64"); |
1336 | | // The possibility that mem allocator throws an exception during memory allocation |
1337 | | // This config is for test usage, be careful when changing it. |
1338 | | DEFINE_mDouble(mem_alloc_fault_probability, "0.0"); |
1339 | | // The time out milliseconds for remote fetch schema RPC, default 60s |
1340 | | DEFINE_mInt64(fetch_remote_schema_rpc_timeout_ms, "60000"); |
1341 | | |
1342 | | DEFINE_Int64(s3_file_system_local_upload_buffer_size, "5242880"); |
1343 | | |
1344 | | //JVM monitoring enable. To prevent be from crashing due to jvm compatibility issues. The default setting is off. |
1345 | | DEFINE_Bool(enable_jvm_monitor, "false"); |
1346 | | |
1347 | | DEFINE_Int32(load_data_dirs_threads, "-1"); |
1348 | | |
1349 | | // Skip loading stale rowset meta when initializing `TabletMeta` from protobuf |
1350 | | DEFINE_mBool(skip_loading_stale_rowset_meta, "false"); |
1351 | | |
1352 | | DEFINE_Bool(enable_file_logger, "true"); |
1353 | | |
1354 | | // The minimum row group size when exporting Parquet files. default 128MB |
1355 | | DEFINE_Int64(min_row_group_size, "134217728"); |
1356 | | |
1357 | | DEFINE_mInt64(compaction_memory_bytes_limit, "1073741824"); |
1358 | | |
1359 | | DEFINE_mInt64(compaction_batch_size, "-1"); |
1360 | | |
1361 | | // If set to false, the parquet reader will not use page index to filter data. |
1362 | | // This is only for debug purpose, in case sometimes the page index |
1363 | | // filter wrong data. |
1364 | | DEFINE_mBool(enable_parquet_page_index, "false"); |
1365 | | |
1366 | | DEFINE_mBool(ignore_not_found_file_in_external_table, "true"); |
1367 | | |
1368 | | DEFINE_mBool(enable_hdfs_mem_limiter, "true"); |
1369 | | |
1370 | | DEFINE_mInt16(topn_agg_limit_multiplier, "2"); |
1371 | | |
1372 | | // Tablet meta size limit after serialization, 1.5GB |
1373 | | DEFINE_mInt64(tablet_meta_serialize_size_limit, "1610612736"); |
1374 | | // Protobuf supports a maximum of 2GB, so the size of the tablet meta after serialization must be less than 2GB |
1375 | | // 1717986918 = 2GB * 0.8 |
1376 | | DEFINE_Validator(tablet_meta_serialize_size_limit, |
1377 | | [](const int64_t config) -> bool { return config < 1717986918; }); |
1378 | | |
1379 | | DEFINE_mInt64(pipeline_task_leakage_detect_period_secs, "60"); |
1380 | | DEFINE_mInt32(snappy_compression_block_size, "262144"); |
1381 | | DEFINE_mInt32(lz4_compression_block_size, "262144"); |
1382 | | |
1383 | | DEFINE_mBool(enable_pipeline_task_leakage_detect, "false"); |
1384 | | |
1385 | | DEFINE_mInt32(check_score_rounds_num, "1000"); |
1386 | | |
1387 | | DEFINE_Int32(query_cache_size, "512"); |
1388 | | |
1389 | | DEFINE_mBool(enable_delete_bitmap_merge_on_compaction, "false"); |
1390 | | // Enable validation to check the correctness of table size. |
1391 | | DEFINE_Bool(enable_table_size_correctness_check, "false"); |
1392 | | DEFINE_Bool(force_regenerate_rowsetid_on_start_error, "false"); |
1393 | | |
1394 | | // clang-format off |
1395 | | #ifdef BE_TEST |
1396 | | // test s3 |
1397 | | DEFINE_String(test_s3_resource, "resource"); |
1398 | | DEFINE_String(test_s3_ak, "ak"); |
1399 | | DEFINE_String(test_s3_sk, "sk"); |
1400 | | DEFINE_String(test_s3_endpoint, "endpoint"); |
1401 | | DEFINE_String(test_s3_region, "region"); |
1402 | | DEFINE_String(test_s3_bucket, "bucket"); |
1403 | | DEFINE_String(test_s3_prefix, "prefix"); |
1404 | | #endif |
1405 | | // clang-format on |
1406 | | |
1407 | | std::map<std::string, Register::Field>* Register::_s_field_map = nullptr; |
1408 | | std::map<std::string, std::function<bool()>>* RegisterConfValidator::_s_field_validator = nullptr; |
1409 | | std::map<std::string, std::string>* full_conf_map = nullptr; |
1410 | | |
1411 | | std::mutex custom_conf_lock; |
1412 | | |
1413 | | std::mutex mutable_string_config_lock; |
1414 | | |
1415 | | // trim string |
1416 | 807 | std::string& trim(std::string& s) { |
1417 | | // rtrim |
1418 | 807 | s.erase(std::find_if(s.rbegin(), s.rend(), [](unsigned char c) { return !std::isspace(c); }) |
1419 | 807 | .base(), |
1420 | 807 | s.end()); |
1421 | | // ltrim |
1422 | 807 | s.erase(s.begin(), |
1423 | 807 | std::find_if(s.begin(), s.end(), [](unsigned char c) { return !std::isspace(c); })); |
1424 | 807 | return s; |
1425 | 807 | } |
1426 | | |
1427 | | // split string by '=' |
1428 | 24 | void splitkv(const std::string& s, std::string& k, std::string& v) { |
1429 | 24 | const char sep = '='; |
1430 | 24 | int start = 0; |
1431 | 24 | int end = 0; |
1432 | 24 | if ((end = s.find(sep, start)) != std::string::npos) { |
1433 | 24 | k = s.substr(start, end - start); |
1434 | 24 | v = s.substr(end + 1); |
1435 | 24 | } else { |
1436 | 0 | k = s; |
1437 | 0 | v = ""; |
1438 | 0 | } |
1439 | 24 | } |
1440 | | |
1441 | | // replace env variables |
1442 | 629 | bool replaceenv(std::string& s) { |
1443 | 629 | std::size_t pos = 0; |
1444 | 629 | std::size_t start = 0; |
1445 | 640 | while ((start = s.find("${", pos)) != std::string::npos) { |
1446 | 11 | std::size_t end = s.find('}', start + 2); |
1447 | 11 | if (end == std::string::npos) { |
1448 | 0 | return false; |
1449 | 0 | } |
1450 | 11 | std::string envkey = s.substr(start + 2, end - start - 2); |
1451 | 11 | const char* envval = std::getenv(envkey.c_str()); |
1452 | 11 | if (envval == nullptr) { |
1453 | 0 | return false; |
1454 | 0 | } |
1455 | 11 | s.erase(start, end - start + 1); |
1456 | 11 | s.insert(start, envval); |
1457 | 11 | pos = start + strlen(envval); |
1458 | 11 | } |
1459 | 629 | return true; |
1460 | 629 | } |
1461 | | |
1462 | | bool strtox(const std::string& valstr, bool& retval); |
1463 | | bool strtox(const std::string& valstr, int16_t& retval); |
1464 | | bool strtox(const std::string& valstr, int32_t& retval); |
1465 | | bool strtox(const std::string& valstr, int64_t& retval); |
1466 | | bool strtox(const std::string& valstr, double& retval); |
1467 | | bool strtox(const std::string& valstr, std::string& retval); |
1468 | | |
1469 | | template <typename T> |
1470 | 7 | bool strtox(const std::string& valstr, std::vector<T>& retval) { |
1471 | 7 | std::stringstream ss(valstr); |
1472 | 7 | std::string item; |
1473 | 7 | T t; |
1474 | 26 | while (std::getline(ss, item, ',')) { |
1475 | 19 | if (!strtox(trim(item), t)) { |
1476 | 0 | return false; |
1477 | 0 | } |
1478 | 19 | retval.push_back(t); |
1479 | 19 | } |
1480 | 7 | return true; |
1481 | 7 | } _ZN5doris6config6strtoxIbEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERSt6vectorIT_SaISB_EE Line | Count | Source | 1470 | 1 | bool strtox(const std::string& valstr, std::vector<T>& retval) { | 1471 | 1 | std::stringstream ss(valstr); | 1472 | 1 | std::string item; | 1473 | 1 | T t; | 1474 | 4 | while (std::getline(ss, item, ',')) { | 1475 | 3 | if (!strtox(trim(item), t)) { | 1476 | 0 | return false; | 1477 | 0 | } | 1478 | 3 | retval.push_back(t); | 1479 | 3 | } | 1480 | 1 | return true; | 1481 | 1 | } |
_ZN5doris6config6strtoxIsEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERSt6vectorIT_SaISB_EE Line | Count | Source | 1470 | 1 | bool strtox(const std::string& valstr, std::vector<T>& retval) { | 1471 | 1 | std::stringstream ss(valstr); | 1472 | 1 | std::string item; | 1473 | 1 | T t; | 1474 | 4 | while (std::getline(ss, item, ',')) { | 1475 | 3 | if (!strtox(trim(item), t)) { | 1476 | 0 | return false; | 1477 | 0 | } | 1478 | 3 | retval.push_back(t); | 1479 | 3 | } | 1480 | 1 | return true; | 1481 | 1 | } |
_ZN5doris6config6strtoxIiEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERSt6vectorIT_SaISB_EE Line | Count | Source | 1470 | 1 | bool strtox(const std::string& valstr, std::vector<T>& retval) { | 1471 | 1 | std::stringstream ss(valstr); | 1472 | 1 | std::string item; | 1473 | 1 | T t; | 1474 | 4 | while (std::getline(ss, item, ',')) { | 1475 | 3 | if (!strtox(trim(item), t)) { | 1476 | 0 | return false; | 1477 | 0 | } | 1478 | 3 | retval.push_back(t); | 1479 | 3 | } | 1480 | 1 | return true; | 1481 | 1 | } |
_ZN5doris6config6strtoxIlEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERSt6vectorIT_SaISB_EE Line | Count | Source | 1470 | 1 | bool strtox(const std::string& valstr, std::vector<T>& retval) { | 1471 | 1 | std::stringstream ss(valstr); | 1472 | 1 | std::string item; | 1473 | 1 | T t; | 1474 | 4 | while (std::getline(ss, item, ',')) { | 1475 | 3 | if (!strtox(trim(item), t)) { | 1476 | 0 | return false; | 1477 | 0 | } | 1478 | 3 | retval.push_back(t); | 1479 | 3 | } | 1480 | 1 | return true; | 1481 | 1 | } |
_ZN5doris6config6strtoxIdEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERSt6vectorIT_SaISB_EE Line | Count | Source | 1470 | 1 | bool strtox(const std::string& valstr, std::vector<T>& retval) { | 1471 | 1 | std::stringstream ss(valstr); | 1472 | 1 | std::string item; | 1473 | 1 | T t; | 1474 | 4 | while (std::getline(ss, item, ',')) { | 1475 | 3 | if (!strtox(trim(item), t)) { | 1476 | 0 | return false; | 1477 | 0 | } | 1478 | 3 | retval.push_back(t); | 1479 | 3 | } | 1480 | 1 | return true; | 1481 | 1 | } |
_ZN5doris6config6strtoxINSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEEEEbRKS7_RSt6vectorIT_SaISB_EE Line | Count | Source | 1470 | 2 | bool strtox(const std::string& valstr, std::vector<T>& retval) { | 1471 | 2 | std::stringstream ss(valstr); | 1472 | 2 | std::string item; | 1473 | 2 | T t; | 1474 | 6 | while (std::getline(ss, item, ',')) { | 1475 | 4 | if (!strtox(trim(item), t)) { | 1476 | 0 | return false; | 1477 | 0 | } | 1478 | 4 | retval.push_back(t); | 1479 | 4 | } | 1480 | 2 | return true; | 1481 | 2 | } |
|
1482 | | |
1483 | 127 | bool strtox(const std::string& valstr, bool& retval) { |
1484 | 127 | if (valstr == "true") { |
1485 | 61 | retval = true; |
1486 | 66 | } else if (valstr == "false") { |
1487 | 65 | retval = false; |
1488 | 65 | } else { |
1489 | 1 | return false; |
1490 | 1 | } |
1491 | 126 | return true; |
1492 | 127 | } |
1493 | | |
1494 | | template <typename T> |
1495 | 419 | bool strtointeger(const std::string& valstr, T& retval) { |
1496 | 419 | if (valstr.length() == 0) { |
1497 | 0 | return false; // empty-string is only allowed for string type. |
1498 | 0 | } |
1499 | 419 | char* end; |
1500 | 419 | errno = 0; |
1501 | 419 | const char* valcstr = valstr.c_str(); |
1502 | 419 | int64_t ret64 = strtoll(valcstr, &end, 10); |
1503 | 419 | if (errno || end != valcstr + strlen(valcstr)) { |
1504 | 0 | return false; // bad parse |
1505 | 0 | } |
1506 | 419 | T tmp = retval; |
1507 | 419 | retval = static_cast<T>(ret64); |
1508 | 419 | if (retval != ret64) { |
1509 | 1 | retval = tmp; |
1510 | 1 | return false; |
1511 | 1 | } |
1512 | 418 | return true; |
1513 | 419 | } _ZN5doris6config12strtointegerIsEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERT_ Line | Count | Source | 1495 | 11 | bool strtointeger(const std::string& valstr, T& retval) { | 1496 | 11 | if (valstr.length() == 0) { | 1497 | 0 | return false; // empty-string is only allowed for string type. | 1498 | 0 | } | 1499 | 11 | char* end; | 1500 | 11 | errno = 0; | 1501 | 11 | const char* valcstr = valstr.c_str(); | 1502 | 11 | int64_t ret64 = strtoll(valcstr, &end, 10); | 1503 | 11 | if (errno || end != valcstr + strlen(valcstr)) { | 1504 | 0 | return false; // bad parse | 1505 | 0 | } | 1506 | 11 | T tmp = retval; | 1507 | 11 | retval = static_cast<T>(ret64); | 1508 | 11 | if (retval != ret64) { | 1509 | 0 | retval = tmp; | 1510 | 0 | return false; | 1511 | 0 | } | 1512 | 11 | return true; | 1513 | 11 | } |
_ZN5doris6config12strtointegerIiEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERT_ Line | Count | Source | 1495 | 301 | bool strtointeger(const std::string& valstr, T& retval) { | 1496 | 301 | if (valstr.length() == 0) { | 1497 | 0 | return false; // empty-string is only allowed for string type. | 1498 | 0 | } | 1499 | 301 | char* end; | 1500 | 301 | errno = 0; | 1501 | 301 | const char* valcstr = valstr.c_str(); | 1502 | 301 | int64_t ret64 = strtoll(valcstr, &end, 10); | 1503 | 301 | if (errno || end != valcstr + strlen(valcstr)) { | 1504 | 0 | return false; // bad parse | 1505 | 0 | } | 1506 | 301 | T tmp = retval; | 1507 | 301 | retval = static_cast<T>(ret64); | 1508 | 301 | if (retval != ret64) { | 1509 | 1 | retval = tmp; | 1510 | 1 | return false; | 1511 | 1 | } | 1512 | 300 | return true; | 1513 | 301 | } |
_ZN5doris6config12strtointegerIlEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERT_ Line | Count | Source | 1495 | 107 | bool strtointeger(const std::string& valstr, T& retval) { | 1496 | 107 | if (valstr.length() == 0) { | 1497 | 0 | return false; // empty-string is only allowed for string type. | 1498 | 0 | } | 1499 | 107 | char* end; | 1500 | 107 | errno = 0; | 1501 | 107 | const char* valcstr = valstr.c_str(); | 1502 | 107 | int64_t ret64 = strtoll(valcstr, &end, 10); | 1503 | 107 | if (errno || end != valcstr + strlen(valcstr)) { | 1504 | 0 | return false; // bad parse | 1505 | 0 | } | 1506 | 107 | T tmp = retval; | 1507 | 107 | retval = static_cast<T>(ret64); | 1508 | 107 | if (retval != ret64) { | 1509 | 0 | retval = tmp; | 1510 | 0 | return false; | 1511 | 0 | } | 1512 | 107 | return true; | 1513 | 107 | } |
|
1514 | | |
1515 | 11 | bool strtox(const std::string& valstr, int16_t& retval) { |
1516 | 11 | return strtointeger(valstr, retval); |
1517 | 11 | } |
1518 | | |
1519 | 301 | bool strtox(const std::string& valstr, int32_t& retval) { |
1520 | 301 | return strtointeger(valstr, retval); |
1521 | 301 | } |
1522 | | |
1523 | 107 | bool strtox(const std::string& valstr, int64_t& retval) { |
1524 | 107 | return strtointeger(valstr, retval); |
1525 | 107 | } |
1526 | | |
1527 | 22 | bool strtox(const std::string& valstr, double& retval) { |
1528 | 22 | if (valstr.length() == 0) { |
1529 | 1 | return false; // empty-string is only allowed for string type. |
1530 | 1 | } |
1531 | 21 | char* end = nullptr; |
1532 | 21 | errno = 0; |
1533 | 21 | const char* valcstr = valstr.c_str(); |
1534 | 21 | retval = strtod(valcstr, &end); |
1535 | 21 | if (errno || end != valcstr + strlen(valcstr)) { |
1536 | 0 | return false; // bad parse |
1537 | 0 | } |
1538 | 21 | return true; |
1539 | 21 | } |
1540 | | |
1541 | 73 | bool strtox(const std::string& valstr, std::string& retval) { |
1542 | 73 | retval = valstr; |
1543 | 73 | return true; |
1544 | 73 | } |
1545 | | |
1546 | | template <typename T> |
1547 | 629 | bool convert(const std::string& value, T& retval) { |
1548 | 629 | std::string valstr(value); |
1549 | 629 | trim(valstr); |
1550 | 629 | if (!replaceenv(valstr)) { |
1551 | 0 | return false; |
1552 | 0 | } |
1553 | 629 | return strtox(valstr, retval); |
1554 | 629 | } _ZN5doris6config7convertISt6vectorIbSaIbEEEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERT_ Line | Count | Source | 1547 | 1 | bool convert(const std::string& value, T& retval) { | 1548 | 1 | std::string valstr(value); | 1549 | 1 | trim(valstr); | 1550 | 1 | if (!replaceenv(valstr)) { | 1551 | 0 | return false; | 1552 | 0 | } | 1553 | 1 | return strtox(valstr, retval); | 1554 | 1 | } |
_ZN5doris6config7convertISt6vectorIsSaIsEEEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERT_ Line | Count | Source | 1547 | 1 | bool convert(const std::string& value, T& retval) { | 1548 | 1 | std::string valstr(value); | 1549 | 1 | trim(valstr); | 1550 | 1 | if (!replaceenv(valstr)) { | 1551 | 0 | return false; | 1552 | 0 | } | 1553 | 1 | return strtox(valstr, retval); | 1554 | 1 | } |
_ZN5doris6config7convertISt6vectorIiSaIiEEEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERT_ Line | Count | Source | 1547 | 1 | bool convert(const std::string& value, T& retval) { | 1548 | 1 | std::string valstr(value); | 1549 | 1 | trim(valstr); | 1550 | 1 | if (!replaceenv(valstr)) { | 1551 | 0 | return false; | 1552 | 0 | } | 1553 | 1 | return strtox(valstr, retval); | 1554 | 1 | } |
_ZN5doris6config7convertISt6vectorIlSaIlEEEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERT_ Line | Count | Source | 1547 | 1 | bool convert(const std::string& value, T& retval) { | 1548 | 1 | std::string valstr(value); | 1549 | 1 | trim(valstr); | 1550 | 1 | if (!replaceenv(valstr)) { | 1551 | 0 | return false; | 1552 | 0 | } | 1553 | 1 | return strtox(valstr, retval); | 1554 | 1 | } |
_ZN5doris6config7convertISt6vectorIdSaIdEEEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERT_ Line | Count | Source | 1547 | 1 | bool convert(const std::string& value, T& retval) { | 1548 | 1 | std::string valstr(value); | 1549 | 1 | trim(valstr); | 1550 | 1 | if (!replaceenv(valstr)) { | 1551 | 0 | return false; | 1552 | 0 | } | 1553 | 1 | return strtox(valstr, retval); | 1554 | 1 | } |
_ZN5doris6config7convertISt6vectorINSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEESaIS8_EEEEbRKS8_RT_ Line | Count | Source | 1547 | 2 | bool convert(const std::string& value, T& retval) { | 1548 | 2 | std::string valstr(value); | 1549 | 2 | trim(valstr); | 1550 | 2 | if (!replaceenv(valstr)) { | 1551 | 0 | return false; | 1552 | 0 | } | 1553 | 2 | return strtox(valstr, retval); | 1554 | 2 | } |
_ZN5doris6config7convertIbEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERT_ Line | Count | Source | 1547 | 124 | bool convert(const std::string& value, T& retval) { | 1548 | 124 | std::string valstr(value); | 1549 | 124 | trim(valstr); | 1550 | 124 | if (!replaceenv(valstr)) { | 1551 | 0 | return false; | 1552 | 0 | } | 1553 | 124 | return strtox(valstr, retval); | 1554 | 124 | } |
_ZN5doris6config7convertIsEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERT_ Line | Count | Source | 1547 | 8 | bool convert(const std::string& value, T& retval) { | 1548 | 8 | std::string valstr(value); | 1549 | 8 | trim(valstr); | 1550 | 8 | if (!replaceenv(valstr)) { | 1551 | 0 | return false; | 1552 | 0 | } | 1553 | 8 | return strtox(valstr, retval); | 1554 | 8 | } |
_ZN5doris6config7convertIiEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERT_ Line | Count | Source | 1547 | 298 | bool convert(const std::string& value, T& retval) { | 1548 | 298 | std::string valstr(value); | 1549 | 298 | trim(valstr); | 1550 | 298 | if (!replaceenv(valstr)) { | 1551 | 0 | return false; | 1552 | 0 | } | 1553 | 298 | return strtox(valstr, retval); | 1554 | 298 | } |
_ZN5doris6config7convertIlEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERT_ Line | Count | Source | 1547 | 104 | bool convert(const std::string& value, T& retval) { | 1548 | 104 | std::string valstr(value); | 1549 | 104 | trim(valstr); | 1550 | 104 | if (!replaceenv(valstr)) { | 1551 | 0 | return false; | 1552 | 0 | } | 1553 | 104 | return strtox(valstr, retval); | 1554 | 104 | } |
_ZN5doris6config7convertIdEEbRKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERT_ Line | Count | Source | 1547 | 19 | bool convert(const std::string& value, T& retval) { | 1548 | 19 | std::string valstr(value); | 1549 | 19 | trim(valstr); | 1550 | 19 | if (!replaceenv(valstr)) { | 1551 | 0 | return false; | 1552 | 0 | } | 1553 | 19 | return strtox(valstr, retval); | 1554 | 19 | } |
_ZN5doris6config7convertINSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEEEEbRKS7_RT_ Line | Count | Source | 1547 | 69 | bool convert(const std::string& value, T& retval) { | 1548 | 69 | std::string valstr(value); | 1549 | 69 | trim(valstr); | 1550 | 69 | if (!replaceenv(valstr)) { | 1551 | 0 | return false; | 1552 | 0 | } | 1553 | 69 | return strtox(valstr, retval); | 1554 | 69 | } |
|
1555 | | |
1556 | | // load conf file |
1557 | 6 | bool Properties::load(const char* conf_file, bool must_exist) { |
1558 | | // if conf_file is null, use the empty props |
1559 | 6 | if (conf_file == nullptr) { |
1560 | 2 | return true; |
1561 | 2 | } |
1562 | | |
1563 | | // open the conf file |
1564 | 4 | std::ifstream input(conf_file); |
1565 | 4 | if (!input.is_open()) { |
1566 | 1 | if (must_exist) { |
1567 | 0 | std::cerr << "config::load() failed to open the file:" << conf_file << std::endl; |
1568 | 0 | return false; |
1569 | 0 | } |
1570 | 1 | return true; |
1571 | 1 | } |
1572 | | |
1573 | | // load properties |
1574 | 3 | std::string line; |
1575 | 3 | std::string key; |
1576 | 3 | std::string value; |
1577 | 3 | line.reserve(512); |
1578 | 114 | while (input) { |
1579 | | // read one line at a time |
1580 | 111 | std::getline(input, line); |
1581 | | |
1582 | | // remove left and right spaces |
1583 | 111 | trim(line); |
1584 | | |
1585 | | // ignore comments |
1586 | 111 | if (line.empty() || line[0] == '#') { |
1587 | 87 | continue; |
1588 | 87 | } |
1589 | | |
1590 | | // read key and value |
1591 | 24 | splitkv(line, key, value); |
1592 | 24 | trim(key); |
1593 | 24 | trim(value); |
1594 | | |
1595 | | // insert into file_conf_map |
1596 | 24 | file_conf_map[key] = value; |
1597 | 24 | } |
1598 | | |
1599 | | // close the conf file |
1600 | 3 | input.close(); |
1601 | | |
1602 | 3 | return true; |
1603 | 4 | } |
1604 | | |
1605 | | template <typename T> |
1606 | | bool Properties::get_or_default(const char* key, const char* defstr, T& retval, bool* is_retval_set, |
1607 | 610 | std::string& rawval) const { |
1608 | 610 | const auto& it = file_conf_map.find(std::string(key)); |
1609 | 610 | std::string valstr; |
1610 | 610 | if (it == file_conf_map.end()) { |
1611 | 595 | if (defstr == nullptr) { |
1612 | | // Not found in conf map, and no default value need to be set, just return |
1613 | 0 | *is_retval_set = false; |
1614 | 0 | return true; |
1615 | 595 | } else { |
1616 | 595 | valstr = std::string(defstr); |
1617 | 595 | } |
1618 | 595 | } else { |
1619 | 15 | valstr = it->second; |
1620 | 15 | } |
1621 | 610 | rawval = valstr; |
1622 | 610 | *is_retval_set = true; |
1623 | 610 | return convert(valstr, retval); |
1624 | 610 | } _ZNK5doris6config10Properties14get_or_defaultIbEEbPKcS4_RT_PbRNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE Line | Count | Source | 1607 | 117 | std::string& rawval) const { | 1608 | 117 | const auto& it = file_conf_map.find(std::string(key)); | 1609 | 117 | std::string valstr; | 1610 | 117 | if (it == file_conf_map.end()) { | 1611 | 116 | if (defstr == nullptr) { | 1612 | | // Not found in conf map, and no default value need to be set, just return | 1613 | 0 | *is_retval_set = false; | 1614 | 0 | return true; | 1615 | 116 | } else { | 1616 | 116 | valstr = std::string(defstr); | 1617 | 116 | } | 1618 | 116 | } else { | 1619 | 1 | valstr = it->second; | 1620 | 1 | } | 1621 | 117 | rawval = valstr; | 1622 | 117 | *is_retval_set = true; | 1623 | 117 | return convert(valstr, retval); | 1624 | 117 | } |
_ZNK5doris6config10Properties14get_or_defaultIsEEbPKcS4_RT_PbRNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE Line | Count | Source | 1607 | 4 | std::string& rawval) const { | 1608 | 4 | const auto& it = file_conf_map.find(std::string(key)); | 1609 | 4 | std::string valstr; | 1610 | 4 | if (it == file_conf_map.end()) { | 1611 | 4 | if (defstr == nullptr) { | 1612 | | // Not found in conf map, and no default value need to be set, just return | 1613 | 0 | *is_retval_set = false; | 1614 | 0 | return true; | 1615 | 4 | } else { | 1616 | 4 | valstr = std::string(defstr); | 1617 | 4 | } | 1618 | 4 | } else { | 1619 | 0 | valstr = it->second; | 1620 | 0 | } | 1621 | 4 | rawval = valstr; | 1622 | 4 | *is_retval_set = true; | 1623 | 4 | return convert(valstr, retval); | 1624 | 4 | } |
_ZNK5doris6config10Properties14get_or_defaultIiEEbPKcS4_RT_PbRNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE Line | Count | Source | 1607 | 296 | std::string& rawval) const { | 1608 | 296 | const auto& it = file_conf_map.find(std::string(key)); | 1609 | 296 | std::string valstr; | 1610 | 296 | if (it == file_conf_map.end()) { | 1611 | 290 | if (defstr == nullptr) { | 1612 | | // Not found in conf map, and no default value need to be set, just return | 1613 | 0 | *is_retval_set = false; | 1614 | 0 | return true; | 1615 | 290 | } else { | 1616 | 290 | valstr = std::string(defstr); | 1617 | 290 | } | 1618 | 290 | } else { | 1619 | 6 | valstr = it->second; | 1620 | 6 | } | 1621 | 296 | rawval = valstr; | 1622 | 296 | *is_retval_set = true; | 1623 | 296 | return convert(valstr, retval); | 1624 | 296 | } |
_ZNK5doris6config10Properties14get_or_defaultIlEEbPKcS4_RT_PbRNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE Line | Count | Source | 1607 | 103 | std::string& rawval) const { | 1608 | 103 | const auto& it = file_conf_map.find(std::string(key)); | 1609 | 103 | std::string valstr; | 1610 | 103 | if (it == file_conf_map.end()) { | 1611 | 103 | if (defstr == nullptr) { | 1612 | | // Not found in conf map, and no default value need to be set, just return | 1613 | 0 | *is_retval_set = false; | 1614 | 0 | return true; | 1615 | 103 | } else { | 1616 | 103 | valstr = std::string(defstr); | 1617 | 103 | } | 1618 | 103 | } else { | 1619 | 0 | valstr = it->second; | 1620 | 0 | } | 1621 | 103 | rawval = valstr; | 1622 | 103 | *is_retval_set = true; | 1623 | 103 | return convert(valstr, retval); | 1624 | 103 | } |
_ZNK5doris6config10Properties14get_or_defaultIdEEbPKcS4_RT_PbRNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE Line | Count | Source | 1607 | 17 | std::string& rawval) const { | 1608 | 17 | const auto& it = file_conf_map.find(std::string(key)); | 1609 | 17 | std::string valstr; | 1610 | 17 | if (it == file_conf_map.end()) { | 1611 | 17 | if (defstr == nullptr) { | 1612 | | // Not found in conf map, and no default value need to be set, just return | 1613 | 0 | *is_retval_set = false; | 1614 | 0 | return true; | 1615 | 17 | } else { | 1616 | 17 | valstr = std::string(defstr); | 1617 | 17 | } | 1618 | 17 | } else { | 1619 | 0 | valstr = it->second; | 1620 | 0 | } | 1621 | 17 | rawval = valstr; | 1622 | 17 | *is_retval_set = true; | 1623 | 17 | return convert(valstr, retval); | 1624 | 17 | } |
_ZNK5doris6config10Properties14get_or_defaultINSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEEEEbPKcSA_RT_PbRS8_ Line | Count | Source | 1607 | 66 | std::string& rawval) const { | 1608 | 66 | const auto& it = file_conf_map.find(std::string(key)); | 1609 | 66 | std::string valstr; | 1610 | 66 | if (it == file_conf_map.end()) { | 1611 | 58 | if (defstr == nullptr) { | 1612 | | // Not found in conf map, and no default value need to be set, just return | 1613 | 0 | *is_retval_set = false; | 1614 | 0 | return true; | 1615 | 58 | } else { | 1616 | 58 | valstr = std::string(defstr); | 1617 | 58 | } | 1618 | 58 | } else { | 1619 | 8 | valstr = it->second; | 1620 | 8 | } | 1621 | 66 | rawval = valstr; | 1622 | 66 | *is_retval_set = true; | 1623 | 66 | return convert(valstr, retval); | 1624 | 66 | } |
_ZNK5doris6config10Properties14get_or_defaultISt6vectorIbSaIbEEEEbPKcS7_RT_PbRNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE Line | Count | Source | 1607 | 1 | std::string& rawval) const { | 1608 | 1 | const auto& it = file_conf_map.find(std::string(key)); | 1609 | 1 | std::string valstr; | 1610 | 1 | if (it == file_conf_map.end()) { | 1611 | 1 | if (defstr == nullptr) { | 1612 | | // Not found in conf map, and no default value need to be set, just return | 1613 | 0 | *is_retval_set = false; | 1614 | 0 | return true; | 1615 | 1 | } else { | 1616 | 1 | valstr = std::string(defstr); | 1617 | 1 | } | 1618 | 1 | } else { | 1619 | 0 | valstr = it->second; | 1620 | 0 | } | 1621 | 1 | rawval = valstr; | 1622 | 1 | *is_retval_set = true; | 1623 | 1 | return convert(valstr, retval); | 1624 | 1 | } |
_ZNK5doris6config10Properties14get_or_defaultISt6vectorIsSaIsEEEEbPKcS7_RT_PbRNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE Line | Count | Source | 1607 | 1 | std::string& rawval) const { | 1608 | 1 | const auto& it = file_conf_map.find(std::string(key)); | 1609 | 1 | std::string valstr; | 1610 | 1 | if (it == file_conf_map.end()) { | 1611 | 1 | if (defstr == nullptr) { | 1612 | | // Not found in conf map, and no default value need to be set, just return | 1613 | 0 | *is_retval_set = false; | 1614 | 0 | return true; | 1615 | 1 | } else { | 1616 | 1 | valstr = std::string(defstr); | 1617 | 1 | } | 1618 | 1 | } else { | 1619 | 0 | valstr = it->second; | 1620 | 0 | } | 1621 | 1 | rawval = valstr; | 1622 | 1 | *is_retval_set = true; | 1623 | 1 | return convert(valstr, retval); | 1624 | 1 | } |
_ZNK5doris6config10Properties14get_or_defaultISt6vectorIiSaIiEEEEbPKcS7_RT_PbRNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE Line | Count | Source | 1607 | 1 | std::string& rawval) const { | 1608 | 1 | const auto& it = file_conf_map.find(std::string(key)); | 1609 | 1 | std::string valstr; | 1610 | 1 | if (it == file_conf_map.end()) { | 1611 | 1 | if (defstr == nullptr) { | 1612 | | // Not found in conf map, and no default value need to be set, just return | 1613 | 0 | *is_retval_set = false; | 1614 | 0 | return true; | 1615 | 1 | } else { | 1616 | 1 | valstr = std::string(defstr); | 1617 | 1 | } | 1618 | 1 | } else { | 1619 | 0 | valstr = it->second; | 1620 | 0 | } | 1621 | 1 | rawval = valstr; | 1622 | 1 | *is_retval_set = true; | 1623 | 1 | return convert(valstr, retval); | 1624 | 1 | } |
_ZNK5doris6config10Properties14get_or_defaultISt6vectorIlSaIlEEEEbPKcS7_RT_PbRNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE Line | Count | Source | 1607 | 1 | std::string& rawval) const { | 1608 | 1 | const auto& it = file_conf_map.find(std::string(key)); | 1609 | 1 | std::string valstr; | 1610 | 1 | if (it == file_conf_map.end()) { | 1611 | 1 | if (defstr == nullptr) { | 1612 | | // Not found in conf map, and no default value need to be set, just return | 1613 | 0 | *is_retval_set = false; | 1614 | 0 | return true; | 1615 | 1 | } else { | 1616 | 1 | valstr = std::string(defstr); | 1617 | 1 | } | 1618 | 1 | } else { | 1619 | 0 | valstr = it->second; | 1620 | 0 | } | 1621 | 1 | rawval = valstr; | 1622 | 1 | *is_retval_set = true; | 1623 | 1 | return convert(valstr, retval); | 1624 | 1 | } |
_ZNK5doris6config10Properties14get_or_defaultISt6vectorIdSaIdEEEEbPKcS7_RT_PbRNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE Line | Count | Source | 1607 | 1 | std::string& rawval) const { | 1608 | 1 | const auto& it = file_conf_map.find(std::string(key)); | 1609 | 1 | std::string valstr; | 1610 | 1 | if (it == file_conf_map.end()) { | 1611 | 1 | if (defstr == nullptr) { | 1612 | | // Not found in conf map, and no default value need to be set, just return | 1613 | 0 | *is_retval_set = false; | 1614 | 0 | return true; | 1615 | 1 | } else { | 1616 | 1 | valstr = std::string(defstr); | 1617 | 1 | } | 1618 | 1 | } else { | 1619 | 0 | valstr = it->second; | 1620 | 0 | } | 1621 | 1 | rawval = valstr; | 1622 | 1 | *is_retval_set = true; | 1623 | 1 | return convert(valstr, retval); | 1624 | 1 | } |
_ZNK5doris6config10Properties14get_or_defaultISt6vectorINSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEESaIS9_EEEEbPKcSD_RT_PbRS9_ Line | Count | Source | 1607 | 2 | std::string& rawval) const { | 1608 | 2 | const auto& it = file_conf_map.find(std::string(key)); | 1609 | 2 | std::string valstr; | 1610 | 2 | if (it == file_conf_map.end()) { | 1611 | 2 | if (defstr == nullptr) { | 1612 | | // Not found in conf map, and no default value need to be set, just return | 1613 | 0 | *is_retval_set = false; | 1614 | 0 | return true; | 1615 | 2 | } else { | 1616 | 2 | valstr = std::string(defstr); | 1617 | 2 | } | 1618 | 2 | } else { | 1619 | 0 | valstr = it->second; | 1620 | 0 | } | 1621 | 2 | rawval = valstr; | 1622 | 2 | *is_retval_set = true; | 1623 | 2 | return convert(valstr, retval); | 1624 | 2 | } |
|
1625 | | |
1626 | 0 | void Properties::set(const std::string& key, const std::string& val) { |
1627 | 0 | file_conf_map.emplace(key, val); |
1628 | 0 | } |
1629 | | |
1630 | 3 | void Properties::set_force(const std::string& key, const std::string& val) { |
1631 | 3 | file_conf_map[key] = val; |
1632 | 3 | } |
1633 | | |
1634 | 3 | Status Properties::dump(const std::string& conffile) { |
1635 | 3 | std::string conffile_tmp = conffile + ".tmp"; |
1636 | 3 | io::FileWriterPtr file_writer; |
1637 | 3 | RETURN_IF_ERROR(io::global_local_filesystem()->create_file(conffile_tmp, &file_writer)); |
1638 | 3 | RETURN_IF_ERROR(file_writer->append("# THIS IS AN AUTO GENERATED CONFIG FILE.\n")); |
1639 | 3 | RETURN_IF_ERROR(file_writer->append( |
1640 | 3 | "# You can modify this file manually, and the configurations in this file\n")); |
1641 | 3 | RETURN_IF_ERROR(file_writer->append("# will overwrite the configurations in be.conf\n\n")); |
1642 | | |
1643 | 3 | for (auto const& iter : file_conf_map) { |
1644 | 3 | RETURN_IF_ERROR(file_writer->append(iter.first)); |
1645 | 3 | RETURN_IF_ERROR(file_writer->append(" = ")); |
1646 | 3 | RETURN_IF_ERROR(file_writer->append(iter.second)); |
1647 | 3 | RETURN_IF_ERROR(file_writer->append("\n")); |
1648 | 3 | } |
1649 | | |
1650 | 3 | RETURN_IF_ERROR(file_writer->close()); |
1651 | | |
1652 | 3 | return io::global_local_filesystem()->rename(conffile_tmp, conffile); |
1653 | 3 | } |
1654 | | |
1655 | | template <typename T> |
1656 | 6 | std::ostream& operator<<(std::ostream& out, const std::vector<T>& v) { |
1657 | 6 | size_t last = v.size() - 1; |
1658 | 25 | for (size_t i = 0; i < v.size(); ++i) { |
1659 | 19 | out << v[i]; |
1660 | 19 | if (i != last) { |
1661 | 13 | out << ", "; |
1662 | 13 | } |
1663 | 19 | } |
1664 | 6 | return out; |
1665 | 6 | } _ZN5doris6configlsIbEERSoS2_RKSt6vectorIT_SaIS4_EE Line | Count | Source | 1656 | 1 | std::ostream& operator<<(std::ostream& out, const std::vector<T>& v) { | 1657 | 1 | size_t last = v.size() - 1; | 1658 | 4 | for (size_t i = 0; i < v.size(); ++i) { | 1659 | 3 | out << v[i]; | 1660 | 3 | if (i != last) { | 1661 | 2 | out << ", "; | 1662 | 2 | } | 1663 | 3 | } | 1664 | 1 | return out; | 1665 | 1 | } |
_ZN5doris6configlsIsEERSoS2_RKSt6vectorIT_SaIS4_EE Line | Count | Source | 1656 | 1 | std::ostream& operator<<(std::ostream& out, const std::vector<T>& v) { | 1657 | 1 | size_t last = v.size() - 1; | 1658 | 4 | for (size_t i = 0; i < v.size(); ++i) { | 1659 | 3 | out << v[i]; | 1660 | 3 | if (i != last) { | 1661 | 2 | out << ", "; | 1662 | 2 | } | 1663 | 3 | } | 1664 | 1 | return out; | 1665 | 1 | } |
_ZN5doris6configlsIiEERSoS2_RKSt6vectorIT_SaIS4_EE Line | Count | Source | 1656 | 1 | std::ostream& operator<<(std::ostream& out, const std::vector<T>& v) { | 1657 | 1 | size_t last = v.size() - 1; | 1658 | 4 | for (size_t i = 0; i < v.size(); ++i) { | 1659 | 3 | out << v[i]; | 1660 | 3 | if (i != last) { | 1661 | 2 | out << ", "; | 1662 | 2 | } | 1663 | 3 | } | 1664 | 1 | return out; | 1665 | 1 | } |
_ZN5doris6configlsIlEERSoS2_RKSt6vectorIT_SaIS4_EE Line | Count | Source | 1656 | 1 | std::ostream& operator<<(std::ostream& out, const std::vector<T>& v) { | 1657 | 1 | size_t last = v.size() - 1; | 1658 | 4 | for (size_t i = 0; i < v.size(); ++i) { | 1659 | 3 | out << v[i]; | 1660 | 3 | if (i != last) { | 1661 | 2 | out << ", "; | 1662 | 2 | } | 1663 | 3 | } | 1664 | 1 | return out; | 1665 | 1 | } |
_ZN5doris6configlsIdEERSoS2_RKSt6vectorIT_SaIS4_EE Line | Count | Source | 1656 | 1 | std::ostream& operator<<(std::ostream& out, const std::vector<T>& v) { | 1657 | 1 | size_t last = v.size() - 1; | 1658 | 4 | for (size_t i = 0; i < v.size(); ++i) { | 1659 | 3 | out << v[i]; | 1660 | 3 | if (i != last) { | 1661 | 2 | out << ", "; | 1662 | 2 | } | 1663 | 3 | } | 1664 | 1 | return out; | 1665 | 1 | } |
_ZN5doris6configlsINSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEEEERSoS8_RKSt6vectorIT_SaISA_EE Line | Count | Source | 1656 | 1 | std::ostream& operator<<(std::ostream& out, const std::vector<T>& v) { | 1657 | 1 | size_t last = v.size() - 1; | 1658 | 5 | for (size_t i = 0; i < v.size(); ++i) { | 1659 | 4 | out << v[i]; | 1660 | 4 | if (i != last) { | 1661 | 3 | out << ", "; | 1662 | 3 | } | 1663 | 4 | } | 1664 | 1 | return out; | 1665 | 1 | } |
|
1666 | | |
1667 | | #define SET_FIELD(FIELD, TYPE, FILL_CONF_MAP, SET_TO_DEFAULT) \ |
1668 | 1.97k | if (strcmp((FIELD).type, #TYPE) == 0) { \ |
1669 | 610 | TYPE new_value = TYPE(); \ |
1670 | 610 | bool is_newval_set = false; \ |
1671 | 610 | std::string raw_value; \ |
1672 | 610 | if (!props.get_or_default((FIELD).name, ((SET_TO_DEFAULT) ? (FIELD).defval : nullptr), \ |
1673 | 610 | new_value, &is_newval_set, raw_value)) { \ |
1674 | 0 | std::cerr << "config field error: " << (FIELD).name << " = \"" << raw_value << '"' \ |
1675 | 0 | << std::endl; \ |
1676 | 0 | return false; \ |
1677 | 0 | } \ |
1678 | 610 | if (!is_newval_set) { \ |
1679 | 0 | continue; \ |
1680 | 0 | } \ |
1681 | 610 | TYPE& ref_conf_value = *reinterpret_cast<TYPE*>((FIELD).storage); \ |
1682 | 610 | TYPE old_value = ref_conf_value; \ |
1683 | 610 | ref_conf_value = new_value; \ |
1684 | 610 | if (RegisterConfValidator::_s_field_validator != nullptr) { \ |
1685 | 610 | auto validator = RegisterConfValidator::_s_field_validator->find((FIELD).name); \ |
1686 | 610 | if (validator != RegisterConfValidator::_s_field_validator->end() && \ |
1687 | 610 | !(validator->second)()) { \ |
1688 | 0 | ref_conf_value = old_value; \ |
1689 | 0 | std::cerr << "validate " << (FIELD).name << "=" << new_value << " failed" \ |
1690 | 0 | << std::endl; \ |
1691 | 0 | return false; \ |
1692 | 0 | } \ |
1693 | 610 | } \ |
1694 | 610 | if (FILL_CONF_MAP) { \ |
1695 | 20 | std::ostringstream oss; \ |
1696 | 20 | oss << ref_conf_value; \ |
1697 | 20 | (*full_conf_map)[(FIELD).name] = oss.str(); \ |
1698 | 20 | } \ |
1699 | 610 | continue; \ |
1700 | 610 | } |
1701 | | |
1702 | | // init conf fields |
1703 | 3 | bool init(const char* conf_file, bool fill_conf_map, bool must_exist, bool set_to_default) { |
1704 | 3 | Properties props; |
1705 | | // load properties file |
1706 | 3 | if (!props.load(conf_file, must_exist)) { |
1707 | 0 | return false; |
1708 | 0 | } |
1709 | | // fill full_conf_map ? |
1710 | 3 | if (fill_conf_map && full_conf_map == nullptr) { |
1711 | 1 | full_conf_map = new std::map<std::string, std::string>(); |
1712 | 1 | } |
1713 | | |
1714 | | // set conf fields |
1715 | 610 | for (const auto& it : *Register::_s_field_map) { |
1716 | 610 | SET_FIELD(it.second, bool, fill_conf_map, set_to_default); |
1717 | 493 | SET_FIELD(it.second, int16_t, fill_conf_map, set_to_default); |
1718 | 489 | SET_FIELD(it.second, int32_t, fill_conf_map, set_to_default); |
1719 | 193 | SET_FIELD(it.second, int64_t, fill_conf_map, set_to_default); |
1720 | 90 | SET_FIELD(it.second, double, fill_conf_map, set_to_default); |
1721 | 73 | SET_FIELD(it.second, std::string, fill_conf_map, set_to_default); |
1722 | 7 | SET_FIELD(it.second, std::vector<bool>, fill_conf_map, set_to_default); |
1723 | 6 | SET_FIELD(it.second, std::vector<int16_t>, fill_conf_map, set_to_default); |
1724 | 5 | SET_FIELD(it.second, std::vector<int32_t>, fill_conf_map, set_to_default); |
1725 | 4 | SET_FIELD(it.second, std::vector<int64_t>, fill_conf_map, set_to_default); |
1726 | 3 | SET_FIELD(it.second, std::vector<double>, fill_conf_map, set_to_default); |
1727 | 2 | SET_FIELD(it.second, std::vector<std::string>, fill_conf_map, set_to_default); |
1728 | 0 | } |
1729 | | |
1730 | 3 | if (config::is_cloud_mode()) { |
1731 | 0 | auto st = config::set_config("enable_file_cache", "true", true, true); |
1732 | 0 | LOG(INFO) << "set config enable_file_cache " |
1733 | 0 | << "true" |
1734 | 0 | << " " << st; |
1735 | 0 | } |
1736 | | |
1737 | 3 | return true; |
1738 | 3 | } |
1739 | | |
1740 | | #define UPDATE_FIELD(FIELD, VALUE, TYPE, PERSIST) \ |
1741 | 53 | if (strcmp((FIELD).type, #TYPE) == 0) { \ |
1742 | 19 | TYPE new_value; \ |
1743 | 19 | if (!convert((VALUE), new_value)) { \ |
1744 | 3 | return Status::Error<ErrorCode::INVALID_ARGUMENT, false>("convert '{}' as {} failed", \ |
1745 | 3 | VALUE, #TYPE); \ |
1746 | 3 | } \ |
1747 | 19 | TYPE& ref_conf_value = *reinterpret_cast<TYPE*>((FIELD).storage); \ |
1748 | 16 | TYPE old_value = ref_conf_value; \ |
1749 | 16 | if (RegisterConfValidator::_s_field_validator != nullptr) { \ |
1750 | 16 | auto validator = RegisterConfValidator::_s_field_validator->find((FIELD).name); \ |
1751 | 16 | if (validator != RegisterConfValidator::_s_field_validator->end() && \ |
1752 | 16 | !(validator->second)()) { \ |
1753 | 0 | ref_conf_value = old_value; \ |
1754 | 0 | return Status::Error<ErrorCode::INVALID_ARGUMENT, false>("validate {}={} failed", \ |
1755 | 0 | (FIELD).name, new_value); \ |
1756 | 0 | } \ |
1757 | 16 | } \ |
1758 | 16 | ref_conf_value = new_value; \ |
1759 | 16 | if (full_conf_map != nullptr) { \ |
1760 | 15 | std::ostringstream oss; \ |
1761 | 15 | oss << new_value; \ |
1762 | 15 | (*full_conf_map)[(FIELD).name] = oss.str(); \ |
1763 | 15 | } \ |
1764 | 16 | if (PERSIST) { \ |
1765 | 3 | RETURN_IF_ERROR(persist_config(std::string((FIELD).name), VALUE)); \ |
1766 | 3 | } \ |
1767 | 16 | update_config(std::string((FIELD).name), VALUE); \ |
1768 | 16 | return Status::OK(); \ |
1769 | 16 | } |
1770 | | |
1771 | | // write config to be_custom.conf |
1772 | | // the caller need to make sure that the given config is valid |
1773 | 3 | Status persist_config(const std::string& field, const std::string& value) { |
1774 | | // lock to make sure only one thread can modify the be_custom.conf |
1775 | 3 | std::lock_guard<std::mutex> l(custom_conf_lock); |
1776 | | |
1777 | 3 | static const std::string conffile = config::custom_config_dir + "/be_custom.conf"; |
1778 | | |
1779 | 3 | Properties tmp_props; |
1780 | 3 | if (!tmp_props.load(conffile.c_str(), false)) { |
1781 | 0 | LOG(WARNING) << "failed to load " << conffile; |
1782 | 0 | return Status::InternalError("failed to load conf file: {}", conffile); |
1783 | 0 | } |
1784 | | |
1785 | 3 | tmp_props.set_force(field, value); |
1786 | 3 | return tmp_props.dump(conffile); |
1787 | 3 | } |
1788 | | |
1789 | | Status set_config(const std::string& field, const std::string& value, bool need_persist, |
1790 | 22 | bool force) { |
1791 | 22 | auto it = Register::_s_field_map->find(field); |
1792 | 22 | if (it == Register::_s_field_map->end()) { |
1793 | 1 | return Status::Error<ErrorCode::NOT_FOUND, false>("'{}' is not found", field); |
1794 | 1 | } |
1795 | | |
1796 | 21 | if (!force && !it->second.valmutable) { |
1797 | 2 | return Status::Error<ErrorCode::NOT_IMPLEMENTED_ERROR, false>( |
1798 | 2 | "'{}' is not support to modify", field); |
1799 | 2 | } |
1800 | | |
1801 | 19 | UPDATE_FIELD(it->second, value, bool, need_persist); |
1802 | 12 | UPDATE_FIELD(it->second, value, int16_t, need_persist); |
1803 | 8 | UPDATE_FIELD(it->second, value, int32_t, need_persist); |
1804 | 6 | UPDATE_FIELD(it->second, value, int64_t, need_persist); |
1805 | 5 | UPDATE_FIELD(it->second, value, double, need_persist); |
1806 | 3 | { |
1807 | | // add lock to ensure thread safe |
1808 | 3 | std::lock_guard<std::mutex> lock(mutable_string_config_lock); |
1809 | 3 | UPDATE_FIELD(it->second, value, std::string, need_persist); |
1810 | 0 | } |
1811 | | |
1812 | | // The other types are not thread safe to change dynamically. |
1813 | 0 | return Status::Error<ErrorCode::NOT_IMPLEMENTED_ERROR, false>( |
1814 | 0 | "'{}' is type of '{}' which is not support to modify", field, it->second.type); |
1815 | 3 | } |
1816 | | |
1817 | 16 | void update_config(const std::string& field, const std::string& value) { |
1818 | 16 | if ("sys_log_level" == field) { |
1819 | | // update log level |
1820 | 0 | update_logging(field, value); |
1821 | 0 | } |
1822 | 16 | } |
1823 | | |
1824 | 0 | Status set_fuzzy_configs() { |
1825 | 0 | std::unordered_map<std::string, std::string> fuzzy_field_and_value; |
1826 | 0 | std::shared_ptr<std::mt19937_64> generator(new std::mt19937_64()); |
1827 | 0 | generator->seed(std::random_device()()); |
1828 | 0 | std::uniform_int_distribution<int64_t> distribution(0, 100); |
1829 | | |
1830 | | // if have set enable_fuzzy_mode=true in be.conf, will fuzzy those field and values |
1831 | 0 | fuzzy_field_and_value["disable_storage_page_cache"] = |
1832 | 0 | ((distribution(*generator) % 2) == 0) ? "true" : "false"; |
1833 | 0 | fuzzy_field_and_value["enable_system_metrics"] = |
1834 | 0 | ((distribution(*generator) % 2) == 0) ? "true" : "false"; |
1835 | 0 | fuzzy_field_and_value["enable_set_in_bitmap_value"] = |
1836 | 0 | ((distribution(*generator) % 2) == 0) ? "true" : "false"; |
1837 | 0 | fuzzy_field_and_value["enable_shrink_memory"] = |
1838 | 0 | ((distribution(*generator) % 2) == 0) ? "true" : "false"; |
1839 | 0 | fuzzy_field_and_value["string_overflow_size"] = |
1840 | 0 | ((distribution(*generator) % 2) == 0) ? "10" : "4294967295"; |
1841 | |
|
1842 | 0 | fmt::memory_buffer buf; |
1843 | 0 | for (auto& it : fuzzy_field_and_value) { |
1844 | 0 | const auto& field = it.first; |
1845 | 0 | const auto& value = it.second; |
1846 | 0 | RETURN_IF_ERROR(set_config(field, value, false, true)); |
1847 | 0 | fmt::format_to(buf, "{}={}, ", field, value); |
1848 | 0 | } |
1849 | 0 | LOG(INFO) << fmt::format("FUZZY MODE IN BE: those variables have been changed: ({}).", |
1850 | 0 | fmt::to_string(buf)); |
1851 | 0 | return Status::OK(); |
1852 | 0 | } |
1853 | | |
1854 | 0 | std::mutex* get_mutable_string_config_lock() { |
1855 | 0 | return &mutable_string_config_lock; |
1856 | 0 | } |
1857 | | |
1858 | 0 | std::vector<std::vector<std::string>> get_config_info() { |
1859 | 0 | std::vector<std::vector<std::string>> configs; |
1860 | 0 | std::lock_guard<std::mutex> lock(mutable_string_config_lock); |
1861 | 0 | for (const auto& it : *full_conf_map) { |
1862 | 0 | auto field_it = Register::_s_field_map->find(it.first); |
1863 | 0 | if (field_it == Register::_s_field_map->end()) { |
1864 | 0 | continue; |
1865 | 0 | } |
1866 | | |
1867 | 0 | std::vector<std::string> _config; |
1868 | 0 | _config.push_back(it.first); |
1869 | |
|
1870 | 0 | std::string config_val = it.second; |
1871 | | // For compatibility, this PR #32933 change the log dir's config logic, |
1872 | | // and deprecate the `sys_log_dir` config. |
1873 | 0 | if (it.first == "sys_log_dir" && config_val == "") { |
1874 | 0 | config_val = fmt::format("{}/log", std::getenv("DORIS_HOME")); |
1875 | 0 | } |
1876 | |
|
1877 | 0 | _config.emplace_back(field_it->second.type); |
1878 | 0 | if (0 == strcmp(field_it->second.type, "bool")) { |
1879 | 0 | _config.emplace_back(config_val == "1" ? "true" : "false"); |
1880 | 0 | } else { |
1881 | 0 | _config.push_back(config_val); |
1882 | 0 | } |
1883 | 0 | _config.emplace_back(field_it->second.valmutable ? "true" : "false"); |
1884 | |
|
1885 | 0 | configs.push_back(_config); |
1886 | 0 | } |
1887 | 0 | return configs; |
1888 | 0 | } |
1889 | | |
1890 | | } // namespace doris::config |