1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
|
// -*- mode:C++; tab-width:8; c-basic-offset:2; indent-tabs-mode:t -*-
// vim: ts=8 sw=2 smarttab ft=cpp
/*
* Ceph - scalable distributed file system
*
* Copyright (C) 2020 Red Hat, Inc.
*
* This is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License version 2.1, as published by the Free Software
* Foundation. See file COPYING.
*
*/
#include <fmt/format.h>
#include "common/errno.h"
#include "rgw_sync_checkpoint.h"
#include "rgw_sal_rados.h"
#include "rgw_bucket_sync.h"
#include "rgw_data_sync.h"
#include "rgw_http_errors.h"
#include "cls/rgw/cls_rgw_client.h"
#include "services/svc_sys_obj.h"
#include "services/svc_zone.h"
#include "rgw_zone.h"
#define dout_subsys ceph_subsys_rgw
namespace {
std::string incremental_marker(const rgw_bucket_shard_sync_info& info)
{
return BucketIndexShardsManager::get_shard_marker(info.inc_marker.position);
}
bool operator<(const std::vector<rgw_bucket_shard_sync_info>& lhs,
const BucketIndexShardsManager& rhs)
{
for (size_t i = 0; i < lhs.size(); ++i) {
const auto& l = incremental_marker(lhs[i]);
const auto& r = rhs.get(i, "");
if (l < r) {
return true;
}
}
return false;
}
bool empty(const BucketIndexShardsManager& markers, int size)
{
for (int i = 0; i < size; ++i) {
const auto& m = markers.get(i, "");
if (!m.empty()) {
return false;
}
}
return true;
}
std::ostream& operator<<(std::ostream& out, const std::vector<rgw_bucket_shard_sync_info>& rhs)
{
const char* separator = ""; // first entry has no comma
out << '[';
for (auto& i : rhs) {
out << std::exchange(separator, ", ") << incremental_marker(i);
}
return out << ']';
}
std::ostream& operator<<(std::ostream& out, const BucketIndexShardsManager& rhs)
{
out << '[';
const char* separator = ""; // first entry has no comma
for (auto& [i, marker] : rhs.get()) {
out << std::exchange(separator, ", ") << marker;
}
return out << ']';
}
int bucket_source_sync_checkpoint(const DoutPrefixProvider* dpp,
rgw::sal::RadosStore* store,
const RGWBucketInfo& bucket_info,
const RGWBucketInfo& source_bucket_info,
const rgw_sync_bucket_pipe& pipe,
uint64_t latest_gen,
const BucketIndexShardsManager& remote_markers,
ceph::timespan retry_delay,
ceph::coarse_mono_time timeout_at)
{
const int num_shards = remote_markers.get().size();
rgw_bucket_sync_status full_status;
int r = rgw_read_bucket_full_sync_status(dpp, store, pipe, &full_status, null_yield);
if (r < 0 && r != -ENOENT) { // retry on ENOENT
return r;
}
// wait for incremental
while (full_status.state != BucketSyncState::Incremental) {
const auto delay_until = ceph::coarse_mono_clock::now() + retry_delay;
if (delay_until > timeout_at) {
lderr(store->ctx()) << "bucket checkpoint timed out waiting to reach incremental sync" << dendl;
return -ETIMEDOUT;
}
ldout(store->ctx(), 1) << "waiting to reach incremental sync.." << dendl;
std::this_thread::sleep_until(delay_until);
r = rgw_read_bucket_full_sync_status(dpp, store, pipe, &full_status, null_yield);
if (r < 0 && r != -ENOENT) { // retry on ENOENT
return r;
}
}
// wait for latest_gen
while (full_status.incremental_gen < latest_gen) {
const auto delay_until = ceph::coarse_mono_clock::now() + retry_delay;
if (delay_until > timeout_at) {
lderr(store->ctx()) << "bucket checkpoint timed out waiting to reach "
"latest generation " << latest_gen << dendl;
return -ETIMEDOUT;
}
ldout(store->ctx(), 1) << "waiting to reach latest gen " << latest_gen
<< ", on " << full_status.incremental_gen << ".." << dendl;
std::this_thread::sleep_until(delay_until);
r = rgw_read_bucket_full_sync_status(dpp, store, pipe, &full_status, null_yield);
if (r < 0 && r != -ENOENT) { // retry on ENOENT
return r;
}
}
if (full_status.incremental_gen > latest_gen) {
ldpp_dout(dpp, 1) << "bucket sync caught up with source:\n"
<< " local gen: " << full_status.incremental_gen << '\n'
<< " remote gen: " << latest_gen << dendl;
return 0;
}
if (empty(remote_markers, num_shards)) {
ldpp_dout(dpp, 1) << "bucket sync caught up with empty source" << dendl;
return 0;
}
std::vector<rgw_bucket_shard_sync_info> status;
status.resize(std::max<size_t>(1, num_shards));
r = rgw_read_bucket_inc_sync_status(dpp, store, pipe,
full_status.incremental_gen, &status);
if (r < 0) {
return r;
}
while (status < remote_markers) {
const auto delay_until = ceph::coarse_mono_clock::now() + retry_delay;
if (delay_until > timeout_at) {
ldpp_dout(dpp, 0) << "bucket checkpoint timed out waiting for incremental sync to catch up" << dendl;
return -ETIMEDOUT;
}
ldpp_dout(dpp, 1) << "waiting for incremental sync to catch up:\n"
<< " local status: " << status << '\n'
<< " remote markers: " << remote_markers << dendl;
std::this_thread::sleep_until(delay_until);
r = rgw_read_bucket_inc_sync_status(dpp, store, pipe,
full_status.incremental_gen, &status);
if (r < 0) {
return r;
}
}
ldpp_dout(dpp, 1) << "bucket sync caught up with source:\n"
<< " local status: " << status << '\n'
<< " remote markers: " << remote_markers << dendl;
return 0;
}
int source_bilog_info(const DoutPrefixProvider *dpp,
RGWSI_Zone* zone_svc,
const rgw_sync_bucket_pipe& pipe,
rgw_bucket_index_marker_info& info,
BucketIndexShardsManager& markers,
optional_yield y)
{
ceph_assert(pipe.source.zone);
auto& zone_conn_map = zone_svc->get_zone_conn_map();
auto conn = zone_conn_map.find(pipe.source.zone->id);
if (conn == zone_conn_map.end()) {
return -EINVAL;
}
return rgw_read_remote_bilog_info(dpp, conn->second, *pipe.source.bucket,
info, markers, y);
}
} // anonymous namespace
int rgw_bucket_sync_checkpoint(const DoutPrefixProvider* dpp,
rgw::sal::RadosStore* store,
const RGWBucketSyncPolicyHandler& policy,
const RGWBucketInfo& info,
std::optional<rgw_zone_id> opt_source_zone,
std::optional<rgw_bucket> opt_source_bucket,
ceph::timespan retry_delay,
ceph::coarse_mono_time timeout_at)
{
struct sync_source_entry {
rgw_sync_bucket_pipe pipe;
uint64_t latest_gen = 0;
BucketIndexShardsManager remote_markers;
RGWBucketInfo source_bucket_info;
};
std::list<sync_source_entry> sources;
// fetch remote markers and bucket info in parallel
boost::asio::io_context ioctx;
for (const auto& [source_zone_id, pipe] : policy.get_all_sources()) {
// filter by source zone/bucket
if (opt_source_zone && *opt_source_zone != *pipe.source.zone) {
continue;
}
if (opt_source_bucket && !opt_source_bucket->match(*pipe.source.bucket)) {
continue;
}
auto& entry = sources.emplace_back();
entry.pipe = pipe;
// fetch remote markers
spawn::spawn(ioctx, [&] (yield_context yield) {
auto y = optional_yield{ioctx, yield};
rgw_bucket_index_marker_info info;
int r = source_bilog_info(dpp, store->svc()->zone, entry.pipe,
info, entry.remote_markers, y);
if (r < 0) {
ldpp_dout(dpp, 0) << "failed to fetch remote bilog markers: "
<< cpp_strerror(r) << dendl;
throw std::system_error(-r, std::system_category());
}
entry.latest_gen = info.latest_gen;
});
// fetch source bucket info
spawn::spawn(ioctx, [&] (yield_context yield) {
auto y = optional_yield{ioctx, yield};
int r = store->getRados()->get_bucket_instance_info(
*entry.pipe.source.bucket, entry.source_bucket_info,
nullptr, nullptr, y, dpp);
if (r < 0) {
ldpp_dout(dpp, 0) << "failed to read source bucket info: "
<< cpp_strerror(r) << dendl;
throw std::system_error(-r, std::system_category());
}
});
}
try {
ioctx.run();
} catch (const std::system_error& e) {
return -e.code().value();
}
// checkpoint each source sequentially
for (const auto& e : sources) {
int r = bucket_source_sync_checkpoint(dpp, store, info, e.source_bucket_info,
e.pipe, e.latest_gen, e.remote_markers,
retry_delay, timeout_at);
if (r < 0) {
ldpp_dout(dpp, 0) << "bucket sync checkpoint failed: " << cpp_strerror(r) << dendl;
return r;
}
}
ldpp_dout(dpp, 0) << "bucket checkpoint complete" << dendl;
return 0;
}
|