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
|
// Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
#include "updateoperation.h"
#include <vespa/document/fieldvalue/document.h>
#include <vespa/document/update/documentupdate.h>
#include <vespa/storageapi/message/bucket.h>
#include <vespa/storageapi/message/persistence.h>
#include <vespa/storage/distributor/distributormetricsset.h>
#include <vespa/storage/distributor/distributor_bucket_space.h>
#include <vespa/vdslib/state/clusterstate.h>
#include <vespa/vespalib/util/stringfmt.h>
#include <vespa/log/log.h>
LOG_SETUP(".distributor.operations.external.update");
using document::BucketSpace;
namespace storage::distributor {
UpdateOperation::UpdateOperation(const DistributorNodeContext& node_ctx,
DistributorStripeOperationContext& op_ctx,
DistributorBucketSpace& bucketSpace,
const std::shared_ptr<api::UpdateCommand>& msg,
std::vector<BucketDatabase::Entry> entries,
UpdateMetricSet& metric)
: Operation(),
_tracker(metric, std::make_shared<api::UpdateReply>(*msg), node_ctx, op_ctx, _cancel_scope),
_msg(msg),
_entries(std::move(entries)),
_new_timestamp(_msg->getTimestamp()),
_is_auto_create_update(_msg->create_if_missing()),
_node_ctx(node_ctx),
_op_ctx(op_ctx),
_bucketSpace(bucketSpace),
_newestTimestampLocation(),
_infoAtSendTime(),
_metrics(metric)
{
}
UpdateOperation::~UpdateOperation() = default;
bool
UpdateOperation::anyStorageNodesAvailable() const
{
const auto& clusterState(_bucketSpace.getClusterState());
const auto storageNodeCount(
clusterState.getNodeCount(lib::NodeType::STORAGE));
for (uint16_t i = 0; i < storageNodeCount; ++i) {
const auto& ns(clusterState.getNodeState(
lib::Node(lib::NodeType::STORAGE, i)));
if (ns.getState() == lib::State::UP
|| ns.getState() == lib::State::RETIRED)
{
return true;
}
}
return false;
}
void
UpdateOperation::onStart(DistributorStripeMessageSender& sender)
{
LOG(debug, "Received UPDATE %s for bucket %" PRIx64,
_msg->getDocumentId().toString().c_str(),
_node_ctx.bucket_id_factory().getBucketId(
_msg->getDocumentId()).getRawId());
// Don't do anything if all nodes are down.
if (!anyStorageNodesAvailable()) {
_tracker.fail(sender,
api::ReturnCode(api::ReturnCode::NOT_CONNECTED,
"Can't store document: No storage nodes "
"available"));
return;
}
if (_entries.empty()) {
document::BucketId bucketId(_node_ctx.bucket_id_factory().getBucketId(_msg->getDocumentId()));
_bucketSpace.getBucketDatabase().getParents(bucketId, _entries);
}
if (_entries.empty()) {
_tracker.fail(sender,
api::ReturnCode(api::ReturnCode::OK,
"No buckets found for given document update"));
return;
}
// An UpdateOperation should only be started iff all replicas are consistent
// with each other, so sampling a single replica should be equal to sampling them all.
assert(_entries[0].getBucketInfo().getNodeCount() > 0); // Empty buckets are not allowed
_infoAtSendTime = _entries[0].getBucketInfo().getNodeRef(0).getBucketInfo();
// FIXME(vekterli): this loop will happily update all replicas in the
// bucket sub-tree, but there is nothing here at all which will fail the
// update if we cannot satisfy a desired replication level (not even for
// n-of-m operations).
for (const auto& entry : _entries) {
LOG(spam, "Found bucket %s", entry.toString().c_str());
const std::vector<uint16_t>& nodes = entry->getNodes();
std::vector<MessageTracker::ToSend> messages;
messages.reserve(nodes.size());
for (uint16_t node : nodes) {
auto command = std::make_shared<api::UpdateCommand>(document::Bucket(_msg->getBucket().getBucketSpace(), entry.getBucketId()),
_msg->getUpdate(), _msg->getTimestamp());
copyMessageSettings(*_msg, *command);
command->setOldTimestamp(_msg->getOldTimestamp());
command->setCondition(_msg->getCondition());
if (_msg->has_cached_create_if_missing()) {
command->set_cached_create_if_missing(_msg->create_if_missing());
}
messages.emplace_back(std::move(command), node);
}
_tracker.queueMessageBatch(std::move(messages));
}
_tracker.flushQueue(sender);
_msg = std::shared_ptr<api::UpdateCommand>();
};
void
UpdateOperation::onReceive(DistributorStripeMessageSender& sender,
const std::shared_ptr<api::StorageReply> & msg)
{
auto& reply = static_cast<api::UpdateReply&>(*msg);
if (msg->getType() == api::MessageType::UPDATE_REPLY) {
uint16_t node = _tracker.handleReply(reply);
if (node != (uint16_t)-1) {
if (reply.getResult().getResult() == api::ReturnCode::OK) {
_results.emplace_back(reply.getBucketId(), reply.getBucketInfo(),
adjusted_received_old_timestamp(reply.getOldTimestamp()), node);
}
if (_tracker.getReply().get()) {
auto& replyToSend = static_cast<api::UpdateReply&>(*_tracker.getReply());
uint64_t oldTs = 0;
uint64_t goodNode = 0;
// Find the highest old timestamp.
for (uint32_t i = 0; i < _results.size(); i++) {
if (_results[i].oldTs > oldTs) {
oldTs = _results[i].oldTs;
goodNode = i;
}
}
replyToSend.setOldTimestamp(oldTs);
for (uint32_t i = 0; i < _results.size(); i++) {
if (_results[i].oldTs < oldTs) {
log_inconsistency_warning(reply, _results[goodNode], _results[i]);
replyToSend.setNodeWithNewestTimestamp(_results[goodNode].nodeId);
_newestTimestampLocation.first = _results[goodNode].bucketId;
_newestTimestampLocation.second = _results[goodNode].nodeId;
_metrics.diverging_timestamp_updates.inc();
break;
}
}
}
_tracker.updateFromReply(sender, reply, node);
}
} else {
_tracker.receiveReply(sender, static_cast<api::BucketInfoReply&>(*msg));
}
}
void
UpdateOperation::log_inconsistency_warning(const api::UpdateReply& reply,
const PreviousDocumentVersion& highest_timestamped_version,
const PreviousDocumentVersion& low_timestamped_version)
{
bool low_ts_node_gc = _op_ctx.has_pending_message(low_timestamped_version.nodeId, reply.getBucket(),
api::MessageType::REMOVELOCATION_ID);
bool high_ts_node_gc = _op_ctx.has_pending_message(highest_timestamped_version.nodeId, reply.getBucket(),
api::MessageType::REMOVELOCATION_ID);
LOG(warning, "Update operation for '%s' in bucket %s updated documents with different timestamps. "
"This should not happen and may indicate undetected replica divergence. "
"Found low ts=%" PRIu64 " on node %u (pending GC: %s), "
"highest ts=%" PRIu64 " on node %u (pending GC: %s)",
reply.getDocumentId().toString().c_str(),
reply.getBucket().toString().c_str(),
low_timestamped_version.oldTs, low_timestamped_version.nodeId, (low_ts_node_gc ? "yes" : "no"),
highest_timestamped_version.oldTs, highest_timestamped_version.nodeId, (high_ts_node_gc ? "yes" : "no"));
LOG(warning, "Bucket info prior to update operation was: %s. After update, "
"info on node %u is %s, info on node %u is %s",
_infoAtSendTime.toString().c_str(),
low_timestamped_version.nodeId, low_timestamped_version.bucketInfo.toString().c_str(),
highest_timestamped_version.nodeId, highest_timestamped_version.bucketInfo.toString().c_str());
}
void
UpdateOperation::onClose(DistributorStripeMessageSender& sender)
{
_tracker.fail(sender, api::ReturnCode(api::ReturnCode::ABORTED, "Process is shutting down"));
}
// The backend behavior of "create-if-missing" updates is to return the timestamp of the
// _new_ update operation if the document was created from scratch. The two-phase update
// operation logic auto-detects unexpected inconsistencies and tries to reconcile
// replicas by forcing document versions to that assumed most likely to preserve the history
// of the document. Normally this is the highest updated timestamp, so to avoid newly created
// replicas from overwriting updates that actually updated existing document versions, treat
// a received timestamp == new timestamp as if it were actually a timestamp of zero.
// This mirrors the received timestamp for regular updates that do not find a matching document.
api::Timestamp UpdateOperation::adjusted_received_old_timestamp(api::Timestamp old_ts_from_node) const {
if (!_is_auto_create_update) {
return old_ts_from_node;
}
return (old_ts_from_node != _new_timestamp) ? old_ts_from_node : api::Timestamp(0);
}
}
|
