Coverage Report

Created: 2026-08-06 08:56

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/client/auth/aws_credential_factory.h"
39
#include "cpp/client/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
std::string hide_access_key(const std::string& ak);
53
54
class S3URI;
55
struct S3ClientConf {
56
    std::string endpoint;
57
    std::string region;
58
    std::string ak;
59
    std::string sk;
60
    std::string token;
61
    // For azure we'd better support the bucket at the first time init azure blob container client
62
    std::string bucket;
63
    io::ObjStorageType provider = io::ObjStorageType::AWS;
64
    int max_connections = -1;
65
    int request_timeout_ms = -1;
66
    int connect_timeout_ms = -1;
67
    bool use_virtual_addressing = true;
68
    // For aws s3, no need to override endpoint
69
    bool need_override_endpoint = true;
70
71
    CredProviderType cred_provider_type = CredProviderType::Default;
72
    std::string role_arn;
73
    std::string external_id;
74
    // True when this client is bound to a Doris internal object storage bucket
75
    // (a storage vault in cloud mode). S3ClientFactory wraps such clients with the
76
    // shared rate limiter; external buckets (S3 load, TVF, external catalogs) are
77
    // returned bare in cloud mode.
78
    bool is_internal_bucket = false;
79
80
    // Full-field identity. get_hash() is only good for picking an unordered_map
81
    // bucket; distinct configurations can collide, so never treat hash equality as
82
    // configuration equality.
83
34.2k
    bool operator==(const S3ClientConf&) const = default;
84
85
1.97k
    uint64_t get_hash() const {
86
1.97k
        uint64_t hash_code = 0;
87
        // Use crc32_hash(ak + sk) hash to prevent swapped AK/SK order from producing same result.
88
1.97k
        hash_code ^= crc32_hash(ak + sk);
89
1.97k
        hash_code ^= crc32_hash(token);
90
1.97k
        hash_code ^= crc32_hash(endpoint);
91
1.97k
        hash_code ^= crc32_hash(region);
92
1.97k
        hash_code ^= crc32_hash(bucket);
93
1.97k
        hash_code ^= max_connections;
94
1.97k
        hash_code ^= request_timeout_ms;
95
1.97k
        hash_code ^= connect_timeout_ms;
96
1.97k
        hash_code ^= use_virtual_addressing;
97
1.97k
        hash_code ^= static_cast<int>(provider);
98
99
1.97k
        hash_code ^= static_cast<int>(cred_provider_type);
100
1.97k
        hash_code ^= crc32_hash(role_arn);
101
1.97k
        hash_code ^= crc32_hash(external_id);
102
1.97k
        hash_code ^= is_internal_bucket;
103
1.97k
        return hash_code;
104
1.97k
    }
105
106
63
    std::string to_string() const {
107
63
        return fmt::format(
108
63
                "(ak={}, token={}, endpoint={}, region={}, bucket={}, max_connections={}, "
109
63
                "request_timeout_ms={}, connect_timeout_ms={}, use_virtual_addressing={}, "
110
63
                "cred_provider_type={},role_arn={}, external_id={}, is_internal_bucket={}",
111
63
                hide_access_key(ak), token.empty() ? "" : "******", endpoint, region, bucket,
112
63
                max_connections, request_timeout_ms, connect_timeout_ms, use_virtual_addressing,
113
63
                cred_provider_type, role_arn, external_id, is_internal_bucket);
114
63
    }
115
};
116
117
struct S3ClientConfHash {
118
1.97k
    size_t operator()(const S3ClientConf& conf) const {
119
1.97k
        return static_cast<size_t>(conf.get_hash());
120
1.97k
    }
121
};
122
123
struct S3Conf {
124
    std::string bucket;
125
    std::string prefix;
126
    S3ClientConf client_conf;
127
128
    bool sse_enabled = false;
129
    static S3Conf get_s3_conf(const cloud::ObjectStoreInfoPB&);
130
    static S3Conf get_s3_conf(const TS3StorageParam&);
131
132
35
    std::string to_string() const {
133
35
        return fmt::format("(bucket={}, prefix={}, client_conf={}, sse_enabled={})", bucket, prefix,
134
35
                           client_conf.to_string(), sse_enabled);
135
35
    }
136
};
137
138
class S3ClientFactory {
139
public:
140
    ~S3ClientFactory();
141
142
    static S3ClientFactory& instance();
143
144
    Result<std::shared_ptr<io::ObjStorageClient>> create(const S3ClientConf& s3_conf);
145
146
    static Status convert_properties_to_s3_conf(const std::map<std::string, std::string>& prop,
147
                                                const S3URI& s3_uri, S3Conf* s3_conf);
148
149
84
    static Aws::Client::ClientConfiguration& getClientConfiguration() {
150
        // The default constructor of ClientConfiguration will do some http call
151
        // such as Aws::Internal::GetEC2MetadataClient and other init operation,
152
        // which is unnecessary.
153
        // So here we use a static instance, and deep copy every time
154
        // to avoid unnecessary operations.
155
84
        static Aws::Client::ClientConfiguration instance;
156
84
        instance.requestTimeoutMs = config::aws_client_request_timeout_ms;
157
84
        return instance;
158
84
    }
159
160
    AwsCredentialResult create_aws_credentials_provider(const S3ClientConf& s3_conf);
161
162
#ifdef BE_TEST
163
    void set_client_creator_for_test(
164
            std::function<std::shared_ptr<io::ObjStorageClient>(const S3ClientConf&)> creator);
165
166
    void clear_client_creator_for_test();
167
#endif
168
169
private:
170
    Result<std::shared_ptr<io::ObjStorageBackend>> _create_s3_backend(const S3ClientConf& s3_conf);
171
    Result<std::shared_ptr<io::ObjStorageBackend>> _create_azure_backend(
172
            const S3ClientConf& s3_conf);
173
    S3ClientFactory();
174
175
    Aws::SDKOptions _aws_options;
176
    std::mutex _lock;
177
    std::unordered_map<S3ClientConf, std::shared_ptr<io::ObjStorageClient>, S3ClientConfHash>
178
            _cache;
179
    std::string _ca_cert_file_path;
180
#ifdef BE_TEST
181
    std::function<std::shared_ptr<io::ObjStorageClient>(const S3ClientConf&)> _test_client_creator;
182
#endif
183
};
184
185
} // end namespace doris