Coverage Report

Created: 2026-07-16 20:57

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
be/src/util/s3_util.h
Line
Count
Source
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
#pragma once
19
20
#include <aws/core/Aws.h>
21
#include <aws/core/client/ClientConfiguration.h>
22
#include <aws/s3/S3Errors.h>
23
#include <bvar/bvar.h>
24
#include <fmt/format.h>
25
#include <gen_cpp/AgentService_types.h>
26
#include <gen_cpp/cloud.pb.h>
27
28
#include <functional>
29
#include <map>
30
#include <memory>
31
#include <mutex>
32
#include <string>
33
#include <unordered_map>
34
35
#include "common/status.h"
36
#include "core/string_ref.h"
37
#include "cpp/aws_common.h"
38
#include "cpp/token_bucket_rate_limiter.h"
39
#include "io/fs/obj_storage_client.h"
40
41
namespace Aws::S3 {
42
class S3Client;
43
} // namespace Aws::S3
44
45
namespace bvar {
46
template <typename T>
47
class Adder;
48
}
49
50
namespace doris {
51
52
namespace s3_bvar {
53
extern bvar::LatencyRecorder s3_get_latency;
54
extern bvar::LatencyRecorder s3_put_latency;
55
extern bvar::LatencyRecorder s3_delete_object_latency;
56
extern bvar::LatencyRecorder s3_delete_objects_latency;
57
extern bvar::LatencyRecorder s3_head_latency;
58
extern bvar::LatencyRecorder s3_multi_part_upload_latency;
59
extern bvar::LatencyRecorder s3_list_latency;
60
extern bvar::LatencyRecorder s3_list_object_versions_latency;
61
extern bvar::LatencyRecorder s3_get_bucket_version_latency;
62
extern bvar::LatencyRecorder s3_copy_object_latency;
63
}; // namespace s3_bvar
64
65
std::string hide_access_key(const std::string& ak);
66
int reset_s3_rate_limiter(S3RateLimitType type, size_t max_speed, size_t max_burst, size_t limit);
67
68
bool acquire_s3_qps_rate_limit(S3RateLimitType type);
69
bool acquire_s3_qps_rate_limit(bool is_internal_bucket, S3RateLimitType type);
70
bool acquire_s3_bytes_rate_limit(S3RateLimitType type, size_t bytes);
71
bool acquire_s3_bytes_rate_limit(bool is_internal_bucket, S3RateLimitType type, size_t bytes);
72
bool acquire_s3_rate_limit(S3RateLimitType type, size_t bytes);
73
bool acquire_s3_rate_limit(bool is_internal_bucket, S3RateLimitType type, size_t bytes);
74
// Refresh GET/PUT QPS and bytes limiters from the current configs and startup CPU count.
75
void refresh_s3_rate_limiters();
76
77
struct S3RateLimiterConfig {
78
    int64_t token_per_second = 0;
79
    int64_t bucket_tokens = 0;
80
    int64_t token_limit = 0;
81
};
82
83
#ifdef BE_TEST
84
void set_s3_rate_limiter_cpu_cores_for_test(int64_t cpu_cores);
85
void reset_s3_rate_limiters_from_config_for_test();
86
S3RateLimiterConfig get_effective_s3_rate_limiter_config(S3RateLimitType type);
87
S3RateLimiterConfig get_effective_s3_bytes_rate_limiter_config(S3RateLimitType type);
88
#endif
89
90
class S3URI;
91
struct S3ClientConf {
92
    std::string endpoint;
93
    std::string region;
94
    std::string ak;
95
    std::string sk;
96
    std::string token;
97
    // For azure we'd better support the bucket at the first time init azure blob container client
98
    std::string bucket;
99
    io::ObjStorageType provider = io::ObjStorageType::AWS;
100
    int max_connections = -1;
101
    int request_timeout_ms = -1;
102
    int connect_timeout_ms = -1;
103
    bool use_virtual_addressing = true;
104
    // For aws s3, no need to override endpoint
105
    bool need_override_endpoint = true;
106
107
    CredProviderType cred_provider_type = CredProviderType::Default;
108
    std::string role_arn;
109
    std::string external_id;
110
    // Whether the client is bound to a Doris internal object storage bucket in cloud mode.
111
    bool is_internal_bucket = false;
112
113
659k
    bool operator==(const S3ClientConf&) const = default;
114
115
38.6k
    uint64_t get_hash() const {
116
38.6k
        uint64_t hash_code = 0;
117
        // Use crc32_hash(ak + sk) hash to prevent swapped AK/SK order from producing same result.
118
38.6k
        hash_code ^= crc32_hash(ak + sk);
119
38.6k
        hash_code ^= crc32_hash(token);
120
38.6k
        hash_code ^= crc32_hash(endpoint);
121
38.6k
        hash_code ^= crc32_hash(region);
122
38.6k
        hash_code ^= crc32_hash(bucket);
123
38.6k
        hash_code ^= max_connections;
124
38.6k
        hash_code ^= request_timeout_ms;
125
38.6k
        hash_code ^= connect_timeout_ms;
126
38.6k
        hash_code ^= use_virtual_addressing;
127
38.6k
        hash_code ^= static_cast<int>(provider);
128
129
38.6k
        hash_code ^= static_cast<int>(cred_provider_type);
130
38.6k
        hash_code ^= crc32_hash(role_arn);
131
38.6k
        hash_code ^= crc32_hash(external_id);
132
38.6k
        hash_code ^= is_internal_bucket;
133
38.6k
        return hash_code;
134
38.6k
    }
135
136
122
    std::string to_string() const {
137
122
        return fmt::format(
138
122
                "(ak={}, token={}, endpoint={}, region={}, bucket={}, max_connections={}, "
139
122
                "request_timeout_ms={}, connect_timeout_ms={}, use_virtual_addressing={}, "
140
122
                "cred_provider_type={},role_arn={}, external_id={}, is_internal_bucket={}",
141
122
                hide_access_key(ak), token, endpoint, region, bucket, max_connections,
142
122
                request_timeout_ms, connect_timeout_ms, use_virtual_addressing, cred_provider_type,
143
122
                role_arn, external_id, is_internal_bucket);
144
122
    }
145
};
146
147
struct S3ClientConfHash {
148
38.6k
    size_t operator()(const S3ClientConf& conf) const {
149
38.6k
        return static_cast<size_t>(conf.get_hash());
150
38.6k
    }
151
};
152
153
struct S3Conf {
154
    std::string bucket;
155
    std::string prefix;
156
    S3ClientConf client_conf;
157
158
    bool sse_enabled = false;
159
    static S3Conf get_s3_conf(const cloud::ObjectStoreInfoPB&);
160
    static S3Conf get_s3_conf(const TS3StorageParam&);
161
162
36
    std::string to_string() const {
163
36
        return fmt::format("(bucket={}, prefix={}, client_conf={}, sse_enabled={})", bucket, prefix,
164
36
                           client_conf.to_string(), sse_enabled);
165
36
    }
166
};
167
168
class S3ClientFactory {
169
public:
170
    ~S3ClientFactory();
171
172
    static S3ClientFactory& instance();
173
174
    std::shared_ptr<io::ObjStorageClient> create(const S3ClientConf& s3_conf);
175
176
    static Status convert_properties_to_s3_conf(const std::map<std::string, std::string>& prop,
177
                                                const S3URI& s3_uri, S3Conf* s3_conf);
178
179
86
    static Aws::Client::ClientConfiguration& getClientConfiguration() {
180
        // The default constructor of ClientConfiguration will do some http call
181
        // such as Aws::Internal::GetEC2MetadataClient and other init operation,
182
        // which is unnecessary.
183
        // So here we use a static instance, and deep copy every time
184
        // to avoid unnecessary operations.
185
86
        static Aws::Client::ClientConfiguration instance;
186
86
        instance.requestTimeoutMs = config::aws_client_request_timeout_ms;
187
86
        return instance;
188
86
    }
189
190
    S3RateLimiterHolder* rate_limiter(S3RateLimitType type);
191
    S3RateLimiterHolder* bytes_rate_limiter(S3RateLimitType type);
192
193
    std::shared_ptr<Aws::Auth::AWSCredentialsProvider> get_aws_credentials_provider(
194
            const S3ClientConf& s3_conf);
195
196
#ifdef BE_TEST
197
    void set_client_creator_for_test(
198
            std::function<std::shared_ptr<io::ObjStorageClient>(const S3ClientConf&)> creator);
199
200
    void clear_client_creator_for_test();
201
#endif
202
203
private:
204
    std::shared_ptr<io::ObjStorageClient> _create_s3_client(const S3ClientConf& s3_conf);
205
    std::shared_ptr<io::ObjStorageClient> _create_azure_client(const S3ClientConf& s3_conf);
206
    std::shared_ptr<Aws::Auth::AWSCredentialsProvider> _get_aws_credentials_provider_v1(
207
            const S3ClientConf& s3_conf);
208
    std::shared_ptr<Aws::Auth::AWSCredentialsProvider> _get_aws_credentials_provider_v2(
209
            const S3ClientConf& s3_conf);
210
    std::shared_ptr<Aws::Auth::AWSCredentialsProvider> _create_credentials_provider(
211
            CredProviderType type);
212
213
    S3ClientFactory();
214
215
    Aws::SDKOptions _aws_options;
216
    std::mutex _lock;
217
    std::unordered_map<S3ClientConf, std::shared_ptr<io::ObjStorageClient>, S3ClientConfHash>
218
            _cache;
219
    std::string _ca_cert_file_path;
220
    std::array<std::unique_ptr<S3RateLimiterHolder>, 2> _rate_limiters;
221
    std::array<std::unique_ptr<S3RateLimiterHolder>, 2> _bytes_rate_limiters;
222
#ifdef BE_TEST
223
    std::function<std::shared_ptr<io::ObjStorageClient>(const S3ClientConf&)> _test_client_creator;
224
#endif
225
};
226
227
} // end namespace doris