Fail with a 403 when indexing a document directly into a follower index.
In order to test this change, I had to move specific assertions into a dedicated class and
disable assertions for that class in the rest qa module. I think that is the right trade off.
If a running shard follow task needs to be restarted and
the remote connection seeds have changed then
a shard follow task currently fails with a fatal error.
The change creates the remote client lazily and adjusts
the errors a shard follow task should retry.
This issue was found in test failures in the recently added
ccr rolling upgrade tests. The reason why this issue occurs
more frequently in the rolling upgrade test is because ccr
is setup in local mode (so remote connection seed will become stale) and
all nodes are restarted, which forces the shard follow tasks to get
restarted at some point during the test. Note that these tests
cannot be enabled yet, because this change will need to be backported
to 6.x first. (otherwise the issue still occurs on non upgraded nodes)
I also changed the RestartIndexFollowingIT to setup remote cluster
via persistent settings and to also restart the leader cluster. This
way what happens during the ccr rolling upgrade qa tests, also happens
in this test.
Relates to #37231
This commit implements a straightforward approach to retention lease
expiration. Namely, we inspect which leases are expired when obtaining
the current leases through the replication tracker. At that moment, we
clean the map that persists the retention leases in memory.
This commit is the first in a series which will culminate with
fully-functional shard history retention leases.
Shard history retention leases are aimed at preventing shard history
consumers from having to fallback to expensive file copy operations if
shard history is not available from a certain point. These consumers
include following indices in cross-cluster replication, and local shard
recoveries. A future consumer will be the changes API.
Further, index lifecycle management requires coordinating with some of
these consumers otherwise it could remove the source before all
consumers have finished reading all operations. The notion of shard
history retention leases that we are introducing here will also be used
to address this problem.
Shard history retention leases are a property of the replication group
managed under the authority of the primary. A shard history retention
lease is a combination of an identifier, a retaining sequence number, a
timestamp indicating when the lease was acquired or renewed, and a
string indicating the source of the lease. Being leases they have a
limited lifespan that will expire if not renewed. The idea of these
leases is that all operations above the minimum of all retaining
sequence numbers will be retained during merges (which would otherwise
clear away operations that are soft deleted). These leases will be
periodically persisted to Lucene and restored during recovery, and
broadcast to replicas under certain circumstances.
This commit is merely putting the basics in place. This first commit
only introduces the concept and integrates their use with the soft
delete retention policy. We add some tests to demonstrate the basic
management is correct, and that the soft delete policy is correctly
influenced by the existence of any retention leases. We make no effort
in this commit to implement any of the following:
- timestamps
- expiration
- persistence to and recovery from Lucene
- handoff during primary relocation
- sharing retention leases with replicas
- exposing leases in shard-level statistics
- integration with cross-cluster replication
These will occur individually in follow-up commits.
In Lucene 8 searches can skip non-competitive hits if the total hit count is not requested.
It is also possible to track the number of hits up to a certain threshold. This is a trade off to speed up searches while still being able to know a lower bound of the total hit count. This change adds the ability to set this threshold directly in the track_total_hits search option. A boolean value (true, false) indicates whether the total hit count should be tracked in the response. When set as an integer this option allows to compute a lower bound of the total hits while preserving the ability to skip non-competitive hits when enough matches have been collected.
Relates #33028
Today the routing of a SourceToParse is assigned in a separate step
after the object is created. We can easily forget to set the routing.
With this commit, the routing must be provided in the constructor of
SourceToParse.
Relates #36921
The AutoFollowCoordinator should be resilient to the fact that the follower
index has already been created and in that case it should only update
the auto follow metadata with the fact that the follower index was created.
Relates to #33007
Currently auto follow stats users are unable to see whether an auto follow
error was recent or old. The new timestamp field will help user distinguish
between old and new errors.
Both index following and auto following should be resilient against missing remote connections.
This happens in the case that they get accidentally removed by a user. When this happens
auto following and index following will retry to continue instead of failing with unrecoverable exceptions.
Both the put follow and put auto follow APIs validate whether the
remote cluster connection. The logic added in this change only exists
in case during the lifetime of a follower index or auto follow pattern
the remote connection gets removed. This retry behavior similar how CCR
deals with authorization errors.
Closes#36667Closes#36255
This commit adds a RemoteClusterAwareRequest interface that allows a
request to specify which remote node it should be routed to. The remote
cluster aware client will attempt to route the request directly to this
node. Otherwise it will send it as a proxy action to eventually end up
on the requested node.
It implements the ccr clean_session action with this client.
This is related to #35975. When the shard restore process is complete,
the index mappings need to be updated to ensure that the data in the
files restores is compatible with the follower mappings. This commit
implements a mapping update as the final step in a shard restore.
Currently if a leader index with soft deletes disabled is auto followed then this index is silently ignored.
This commit changes this behavior to mark these indices as auto followed and report an error, which is visible in auto follow stats. Marking the index as auto follow is important, because otherwise the auto follower will continuously try to auto follow and fail.
Relates to #33007
This commit is related to #36127. It adds a CcrRestoreSourceService to
track Engine.IndexCommitRef need for in-process file restores. When a
follower starts restoring a shard through the CcrRepository it opens a
session with the leader through the PutCcrRestoreSessionAction. The
leader responds to the request by telling the follower what files it
needs to fetch for a restore. This is not yet implemented.
Once, the restore is complete, the follower closes the session with the
DeleteCcrRestoreSessionAction action.
Currently, the CcrRepositoryManger only listens for settings updates
and installs new repositories. It does not install the repositories that
are in the initial settings. This commit, modifies the manager to
install the initial repositories. Additionally, it modifies the ccr
integration test to configure the remote leader node at startup, instead
of using a settings update.
For each remote cluster the auto follow coordinator, starts an auto
follower that checks the remote cluster state and determines whether an
index needs to be auto followed. The time since last auto follow is
reported per remote cluster and gives insight whether the auto follow
process is alive.
Relates to #33007
Originates from #35895
If a primary promotion happens in the test testAddRemoveShardOnLeader, the
max_seq_no_of_updates_or_deletes on a new primary might be higher than the
max_seq_no_of_updates_or_deletes on the replicas or copies of the follower.
Relates #36607
This commit add support for using sequence numbers to power [optimistic concurrency control](http://en.wikipedia.org/wiki/Optimistic_concurrency_control)
in the delete and index transport actions and requests. A follow up will come with adding sequence
numbers to the update and get results.
Relates #36148
Relates #10708
For class fields of type collection whose order is not important
and for which duplicates are not permitted we declare them as `Set`s.
Usually the definition is a `HashSet` but in this case `TreeSet` is used
instead to aid testing.
* Enable parallel restore operations
* Add uuid to restore in progress entries to uniquely identify them
* Adjust restore in progress entries to be a map in cluster state
* Added tests for:
* Parallel restore from two different snapshots
* Parallel restore from a single snapshot to different indices to test uuid identifiers are correctly used by `RestoreService` and routing allocator
* Parallel restore with waiting for completion to test transport actions correctly use uuid identifiers
testFailLeaderReplicaShard periodically fails because we concurrently
index to the leader group and close one of its replicas. If a
replication request hits a closing shard, we will fail that shard;
however, failing a shard is supported by the test framework - this makes
the test fail.
This commit add support to engine operations for resolving and verifying the sequence number and
primary term of the last modification to a document before performing an operation. This is
infrastructure to move our (optimistic concurrency control)[http://en.wikipedia.org/wiki/Optimistic_concurrency_control] API to use sequence numbers instead of internal versioning.
Relates #36148
Relates #10708
Changed AutofollowCoordinator makes use of the wait_for_metadata_version
feature in cluster state API and removed hard coded poll interval.
Originates from #35895
Relates to #33007
The auto follow coordinator keeps track of the UUIDs of indices that it has followed. The index UUID strings need to be cleaned up in the case that these indices are removed in the remote cluster.
Relates to #33007
1. CCR tests work without any changes
2. `testDanglingIndices` require changes the source code (added TODO).
3. `testIndexDeletionWhenNodeRejoins` because it's using just two
nodes, adding the node to exclusions is needed on restart.
4. `testCorruptTranslogTruncationOfReplica` starts dedicated master
one, because otherwise, the cluster does not form, if nodes are stopped
and one node is started back.
5. `testResolvePath` needs TEST cluster, because all nodes are stopped
at the end of the test and it's not possible to perform checks needed
by SUITE cluster.
6. `SnapshotDisruptionIT`. Without changes, the test fails because Zen2
retries snapshot creation as soon as network partition heals. This
results into the race between creating snapshot and test cleanup logic
(deleting index). Zen1 on the
other hand, also schedules retry, but it takes some time after network
partition heals, so cleanup logic executes latter and test passes. The
check that snapshot is eventually created is added to
the end of the test.
Renamed the follow qa modules:
`multi-cluster-downgraded-to-basic-license` to `downgraded-to-basic-license`
`multi-cluster-with-non-compliant-license` to `non-compliant-license`
`multi-cluster-with-security` to `security`
Moved the `chain` module into the `multi-cluster` module and
changed the `multi-cluster` to start 3 clusters.
Followup from #36031
This is related to #35975. It implements a basic restore functionality
for the CcrRepository. When the restore process is kicked off, it
configures the new index as expected for a follower index. This means
that the index has a different uuid, the version is not incremented, and
the Ccr metadata is installed.
When the restore shard method is called, an empty shard is initialized.
In #34474, we added a new assertion to ensure that the
LocalCheckpointTracker is always consistent with Lucene index. However,
we reset LocalCheckpoinTracker in testDedupByPrimaryTerm cause this
assertion to be violated.
This commit removes resetCheckpoint from LocalCheckpointTracker and
rewrites testDedupByPrimaryTerm without resetting the local checkpoint.
Relates #34474
This commit moves back to use explicit dependsOn for test tasks on
check. Not all tasks extending RandomizedTestingTask should be run by
check directly.
This commit changes the format of the `hits.total` in the search response to be an object with
a `value` and a `relation`. The `value` indicates the number of hits that match the query and the
`relation` indicates whether the number is accurate (in which case the relation is equals to `eq`)
or a lower bound of the total (in which case it is equals to `gte`).
This change also adds a parameter called `rest_total_hits_as_int` that can be used in the
search APIs to opt out from this change (retrieve the total hits as a number in the rest response).
Note that currently all search responses are accurate (`track_total_hits: true`) or they don't contain
`hits.total` (`track_total_hits: true`). We'll add a way to get a lower bound of the total hits in a
follow up (to allow numbers to be passed to `track_total_hits`).
Relates #33028
This is a follow-up to #36086. It renames the internal repository
actions to be prefixed by "internal". This allows the system user to
execute the actions.
Additionally, this PR stops casting Client to NodeClient. The client we
have is a NodeClient so executing the actions will be local.
and replaced poll interval setting with a hardcoded poll interval.
The hard coded interval will be removed in a follow up change to make
use of cluster state API's wait_for_metatdata_version.
Before the auto following was bootstrapped from thread pool scheduler,
but now auto followers for new remote clusters are bootstrapped when
a new cluster state is published.
Originates from #35895
Relates to #33007
Closes#35435
- make it easier to add additional testing tasks with the proper configuration and add some where they were missing.
- mute or fix failing tests
- add a check as part of testing conventions to find classes not included in any testing task.
This commit replaces usages of Streamable with Writeable for the
BaseTasksResponse / TransportTasksAction classes and subclasses of
these classes.
Note that where possible response fields were made final.
Relates to #34389
The current response format is:
```
{
"pattern1": {
...
},
"pattern2": {
...
}
}
```
The new format is:
```
{
"patterns": [
{
"name": "pattern1",
"pattern": {
...
}
},
{
"name": "pattern2",
"pattern": {
...
}
}
]
}
```
This format is more structured and more friendly for parsing and generating specs.
This is a breaking change, but it is better to do this now while ccr
is still a beta feature than later.
Follow up from #36049
This commit adds an empty CcrRepository snapshot/restore repository.
When a new cluster is registered in the remote cluster settings, a new
CcrRepository is registered for that cluster.
This is implemented using a new concept of "internal repositories".
RepositoryPlugin now allows implementations to return factories for
"internal repositories". The "internal repositories" are different from
normal repositories in that they cannot be registered through the
external repository api. Additionally, "internal repositories" are local
to a node and are not stored in the cluster state.
The repository will be unregistered if the remote cluster is removed.
It is important that all shards of a given index have the same
`indexCreatedVersionMajor` to Lucene, or eg. merging those shards is going to
be considered illegal. At the moment, we use the latest Lucene version when
creating a shard, which could cause shards to have different created versions
eg. in case of forced allocation. This commit makes sure to reuse the
appropriate Lucene version in order to avoid such issues.
Closes#33826
AutoFollowCoordinator should take into account that after auto following
an index and while updating that a leader index has been followed, that
the auto follow pattern may have been removed via delete auto follow patterns
api.
Also fixed a bug that when a remote cluster connection has been removed,
the auto follow coordinator does not die when it tries get a remote client for
that cluster.
Closes#35480
* Replace Streamable w/ Writeable in BaseTasksRequest and subclasses
This commit replaces usages of Streamable with Writeable for the
BaseTasksRequest / TransportTasksAction classes and subclasses of
these classes.
Relates to #34389
Some tests kill nodes and otherwise it would take 60s by default
for replicas to get allocated and that is longer than we wait
for getting in a green state in tests.
Relates to #35403
Today the default for USE_ZEN2 is false and it is overridden in many places. By
defaulting it to true we can be sure that the only places in which Zen2 does
not work are those in which it is explicitly set to false.
This change adds an extra check that verifies that all primary shards
have been started of an index that is about to be auto followed.
If not all primary shards have been started for an index
then the next auto follow run will try to follow to auto follow
this index again.
Closes#35480
When there is no persistent tasks metadata we could hit a null pointer
exception when executing a follower stats request. This is because we
inspect the persistent tasks metadata. Yet, if no tasks have been
registered, this is null (as opposed to empty). We need to avoid
de-referencing the persistent tasks metadata in this case. That is what
this commit does, and we add a test for this situation.
Currently there is a common NPE in the IndexFollowingIT that does not
indicate the test failing. This is when a cluster state listener is
called and certain index metadata is not yet available.
This commit checks that the metadata is not null before performing the
logic that depends on the metadata.
Removed extending of AbstractComponent and changed logger usage to
explicit declaration. Abstract classes still have logger
declaration using this.getClass() in order to show implementation class
name in its logs.
See #34488
avoid the assertions that check the log files, because that does not work on Windows.
The rest of the test is still useful and should work on Windows CI.
Currently on Windows CI this qa module fails because there is just one test and
that test si ignored if OS is Windows.
Validate remote cluster license as part of put auto follow pattern api call
in addition of validation that when auto follow coordinator starts auto
following indices in the leader cluster.
Also added qa module that tests what happens to ccr after downgrading to basic license.
Existing active follow indices should remain to follow,
but the auto follow feature should not pickup new leader indices.
Adjust list of dynamic index settings that should be replicated
and added a test that verifies whether builtin dynamic index settings
are classified as replicated or non replicated (whitelisted).
This commit uses the index settings version so that a follower can
replicate index settings changes as needed from the leader.
Co-authored-by: Martijn van Groningen <martijn.v.groningen@gmail.com>
An auto follow pattern:
* cannot start with `_`
* cannot contain a `,`
* can be encoded in UTF-8
* the length of UTF-8 encoded bytes is no longer than 255 bytes
In order to start shard follow tasks, the resume follow api already
needs execute N requests to the elected master node.
The pause follow API is also a master node action, which would make
how both APIs execute more consistent.
Error was thrown if leader index had no soft deletes enabled, but it then continued creating the follower index.
The test caught this bug, but very rarely due to timing issue.
Build failure instance:
```
1> [2018-11-05T20:29:38,597][INFO ][o.e.x.c.LocalIndexFollowingIT] [testDoNotCreateFollowerIfLeaderDoesNotHaveSoftDeletes] before test
1> [2018-11-05T20:29:38,599][INFO ][o.e.c.s.ClusterSettings ] [node_s_0] updating [cluster.remote.local.seeds] from [[]] to [["127.0.0.1:9300"]]
1> [2018-11-05T20:29:38,599][INFO ][o.e.c.s.ClusterSettings ] [node_s_0] updating [cluster.remote.local.seeds] from [[]] to [["127.0.0.1:9300"]]
1> [2018-11-05T20:29:38,609][INFO ][o.e.c.m.MetaDataCreateIndexService] [node_s_0] [leader-index] creating index, cause [api], templates [random-soft-deletes-templat
e, one_shard_index_template], shards [2]/[0], mappings []
1> [2018-11-05T20:29:38,628][INFO ][o.e.c.r.a.AllocationService] [node_s_0] Cluster health status changed from [YELLOW] to [GREEN] (reason: [shards started [[leader-
index][0]] ...]).
1> [2018-11-05T20:29:38,660][INFO ][o.e.x.c.a.TransportPutFollowAction] [node_s_0] [follower-index] creating index, cause [ccr_create_and_follow], shards [2]/[0]
1> [2018-11-05T20:29:38,675][INFO ][o.e.c.s.ClusterSettings ] [node_s_0] updating [cluster.remote.local.seeds] from [["127.0.0.1:9300"]] to [[]]
1> [2018-11-05T20:29:38,676][INFO ][o.e.c.s.ClusterSettings ] [node_s_0] updating [cluster.remote.local.seeds] from [["127.0.0.1:9300"]] to [[]]
1> [2018-11-05T20:29:38,678][INFO ][o.e.x.c.LocalIndexFollowingIT] [testDoNotCreateFollowerIfLeaderDoesNotHaveSoftDeletes] after test
1> [2018-11-05T20:29:38,678][INFO ][o.e.x.c.LocalIndexFollowingIT] [testDoNotCreateFollowerIfLeaderDoesNotHaveSoftDeletes] [LocalIndexFollowingIT#testDoNotCreateFoll
owerIfLeaderDoesNotHaveSoftDeletes]: cleaning up after test
1> [2018-11-05T20:29:38,678][INFO ][o.e.c.m.MetaDataDeleteIndexService] [node_s_0] [follower-index/TlWlXp0JSVasju2Kr_hksQ] deleting index
1> [2018-11-05T20:29:38,678][INFO ][o.e.c.m.MetaDataDeleteIndexService] [node_s_0] [leader-index/FQ6EwIWcRAKD8qvOg2eS8g] deleting index
FAILURE 0.23s J0 | LocalIndexFollowingIT.testDoNotCreateFollowerIfLeaderDoesNotHaveSoftDeletes <<< FAILURES!
> Throwable #1: java.lang.AssertionError:
> Expected: <false>
> but: was <true>
> at __randomizedtesting.SeedInfo.seed([7A3C89DA3BCA17DD:65C26CBF6FEF0B39]:0)
> at org.hamcrest.MatcherAssert.assertThat(MatcherAssert.java:20)
> at org.elasticsearch.xpack.ccr.LocalIndexFollowingIT.testDoNotCreateFollowerIfLeaderDoesNotHaveSoftDeletes(LocalIndexFollowingIT.java:83)
> at java.lang.Thread.run(Thread.java:748)
```
Build failure: https://elasticsearch-ci.elastic.co/job/elastic+elasticsearch+6.5+intake/46/console
The suite FollowerFailOverIT is failing because some documents are not
replicated to the follower. Maybe the FollowTask is not working as
expected or the background indexers eat all resources while the follower
cluster is trying to reform after a failover; then CI is not fast enough
to replicate all the indexed docs within 60 seconds (sometimes I see 80k
docs on the leader).
This commit limits the number of documents to be indexed into the leader
index by the background threads so that we can eliminate the latter
case. This change also replaces a docCount assertion with a docIds
assertion so we can have more information if these tests fail again.
Relates #33337
Only the response classes of get auto follow pattern, the follow and stats APIs
were moved away from Streamable. The other APIs use `AcknowledgedResponse`
or `BaseTasksResponse` as response class and
moving that class away from Streamable is a bigger change.
Stop passing `Settings` to `AbstractComponent`'s ctor. This allows us to
stop passing around `Settings` in a *ton* of places. While this change
touches many files, it touches them all in fairly small, mechanical
ways, doing a few things per file:
1. Drop the `super(settings);` line on everything that extends
`AbstractComponent`.
2. Drop the `settings` argument to the ctor if it is no longer used.
3. If the file doesn't use `logger` then drop `extends
AbstractComponent` from it.
4. Clean up all compilation failure caused by the `settings` removal
and drop any now unused `settings` isntances and method arguments.
I've intentionally *not* removed the `settings` argument from a few
files:
1. TransportAction
2. AbstractLifecycleComponent
3. BaseRestHandler
These files don't *need* `settings` either, but this change is large
enough as is.
Relates to #34488
Only the follow stats request couldn't be changed to use Writeable serialization,
because that requires changes in `TransportTasksAction` and `BaseTasksRequest` base classes.
This commit fixes two issues with the CCR API specification:
- remove the CCR stats endpoint, it is not currently implemented
- fix the documentation links
* Changed the auto follow stats to also include follow stats.
* Renamed the auto follow stats api to stats api and changed its url path
from `/_ccr/auto_follow/stats` `/_ccr/stats`.
* Removed `/_ccr/stats` url path for the follow stats api, which makes
the index parameter a required parameter.
* Fixed docs.
Index shard stats for the follower shard are fetched, when a shard follow task is started.
This is needed in order to bootstap the shard follow task with the follower global checkpoint.
Sometimes index shard stats are not available (e.g. during a restart) and
we fail now, while it is very likely that these stats will be available some time later.
This limit is based on the size in bytes of the operations in the write buffer. If this limit is exceeded then no more read operations will be coordinated until the size in bytes of the write buffer has dropped below the configured write buffer size limit.
Renamed existing `max_write_buffer_size` to ``max_write_buffer_count` to indicate that limit is count based.
Closes#34705
We should not create a follower index and abort a follow request if the
leader does not have soft-deletes. Moreover, we also should not
auto-follow an index if it does not have soft-deletes.
With this change, we apply the common test config automatically to all
newly created tasks instead of opting in specifically.
For plugin authors using the plugin externally this means that the
configuration will be applied to their RandomizedTestingTasks as well.
The purpose of the task is to simplify setup and make it easier to
change projects that use the `test` task but actually run integration
tests to use a task called `integTest` for clarity, but also because
we may want to configure and run them differently.
E.x. using different levels of concurrency.
Per #31717 this commit changes the defaults to the following:
Batch size of 5120 ops.
Maximum of 12 concurrent read requests.
Maximum of 9 concurrent write requests.
This is not necessarily our final values but it's good to have these as defaults for the purposes of initial testing.
* Change the `TransportPauseFollowAction` to extend from `TransportMasterNodeAction`
instead of `HandledAction`, this removes a sync cluster state api call.
* Introduced `ResponseHandler` that removes duplicated code in `TransportPauseFollowAction` and
`TransportResumeFollowAction`.
* Changed `PauseFollowAction.Request` to not use `readFrom()`.
Both testFollowIndexAndCloseNode and testFailOverOnFollower failed
because they responded to the FollowTask a TransportService closed
exception which is currently considered as a fatal error. This behavior
is not desirable since a closing node can throw that exception, and we
should retry in that case.
This change adds TransportService closed error to the list of retryable
errors.
Closes#34694
As part of this change the leader index name and leader cluster name are
stored in the CCR metadata in the follow index. The resume follow api
will read that when a resume follow request is executed.
Since #34412 and #34474, a follower must have soft-deletes enabled
to work correctly. This change requires soft-deletes on the follower.
Relates #34412
Relates #34474
* Changed the resource id of auto follow patterns to be a user defined name
instead of being the leader cluster alias name.
* Fail when an unfollowed leader index matches with two or more auto follow patterns.
We should be consistent here. We were already using the casing "Ccr" and
this is the preferred casing for Java class names. This commit adjusts
the names of some classes that were using the casing "CCR" to be "Ccr".
In some of our X-Pack REST tests we have to wait for pending tasks to
complete. We are now needing this functionality in ESRestTestCase for
the docs tests where we run against X-Pack features. This commit moves
the helper method that we have in X-Pack to ESRestTestCase, and removes
duplicate logic from waiting for rollup tasks to complete.
Since #34288, we might hit deadlock if the FollowTask has more fetchers
than writers. This can happen in the following scenario:
Suppose the leader has two operations [seq#0, seq#1]; the FollowTask has
two fetchers and one writer.
1. The FollowTask issues two concurrent fetch requests: {from_seq_no: 0,
num_ops:1} and {from_seq_no: 1, num_ops:1} to read seq#0 and seq#1
respectively.
2. The second request which fetches seq#1 completes before, and then it
triggers a write request containing only seq#1.
3. The primary of a follower fails after it has replicated seq#1 to
replicas.
4. Since the old primary did not respond, the FollowTask issues another
write request containing seq#1 (resend the previous write request).
5. The new primary has seq#1 already; thus it won't replicate seq#1 to
replicas but will wait for the global checkpoint to advance at least
seq#1.
The problem is that the FollowTask has only one writer and that writer
is waiting for seq#0 which won't be delivered until the writer completed.
This PR proposes to replicate existing operations with the old primary
term (instead of the current term) on the follower. In particular, when
the following primary detects that it has processed an process already,
it will look up the term of an existing operation with the same seq_no
in the Lucene index, then rewrite that operation with the old term
before replicating it to the following replicas. This approach is
wait-free but requires soft-deletes on the follower.
Relates #34288
Today we rely on the LocalCheckpointTracker to ensure no duplicate when
enabling optimization using max_seq_no_of_updates. The problem is that
the LocalCheckpointTracker is not fully reloaded when opening an engine
with an out-of-order index commit. Suppose the starting commit has seq#0
and seq#2, then the current LocalCheckpointTracker would return "false"
when asking if seq#2 was processed before although seq#2 in the commit.
This change scans the existing sequence numbers in the starting commit,
then marks these as completed in the LocalCheckpointTracker to ensure
the consistent state between LocalCheckpointTracker and Lucene commit.
This change makes it no longer possible to follow / auto follow without
specifying a leader cluster. If a local index needs to be followed
then `cluster.remote.*.seeds` should point to nodes in the local cluster.
Closes#34258
The `AutoFollowTests` needs to restart the clusters between each tests, because
it is using auto follow stats in assertions. Auto follow stats are only reset
by stopping the elected master node.
Extracted the `testGetOperationsBasedOnGlobalSequenceId()` test to its own test, because it just tests the shard changes api.
* Renamed AutoFollowTests to AutoFollowIT, because it is an integration test.
Renamed ShardChangesIT to IndexFollowingIT, because shard changes it the name
of an internal api and isn't a good name for an integration test.
* move creation of NodeConfigurationSource to a seperate method
* Fixes issues after merge, moved assertSeqNos() and assertSameDocIdsOnShards() methods from ESIntegTestCase to InternalTestCluster, so that ccr tests can use these methods too.
Today we rewrite the operations from the leader with the term of the
following primary because the follower should own its history. The
problem is that a newly promoted primary may re-assign its term to
operations which were replicated to replicas before by the previous
primary. If this happens, some operations with the same seq_no may be
assigned different terms. This is not good for the future optimistic
locking using a combination of seqno and term.
This change ensures that the primary of a follower only processes an
operation if that operation was not processed before. The skipped
operations are guaranteed to be delivered to replicas via either
primary-replica resync or peer-recovery. However, the primary must not
acknowledge until the global checkpoint is at least the highest seqno of
all skipped ops (i.e., they all have been processed on every replica).
Relates #31751
Relates #31113
Also fixed ShardFollowNodeTaskTests to not return ops when responseSize
is empty. Otherwise ops are returned when no ops are expected to be returned.
Co-authored-by: Jason Tedor <jason@tedor.me>
Unfollow should be allowed / disallowed on a per index level instead of
cluster level.
Also renamed `create_follow_index` index privilege to
`manage_follow_index` privilege and include unfollow and close APIs.
This commit modifies the follow stats API response structure to more
clearly highlight meaning of the higher level fields. In particular,
previously the response had a top-level key for each index. Instead, we
nest the indices under an "indices" field which is now an array. The
values in this array are objects containing two fields: "index" which is
the name of the follower index, and "shards" which is an array where
each value in the array is the follower stats for that shard. That is,
we have gone from:
{
"bar": [
{
"shard_id": 0...
}...
]...
}
to
{
"indices": [
{
"index": "bar",
"shards": [
{
"shard_id": 0...
}...
]
}...
}
In the CCR docs we want to refer to the endpoint that returns following
stats as the follow stats API. This commit renames the internal
implementation of this endpoint to reflect this usage.
The follower index shard history UUID will be fetched from the indices stats api when the shard follow task starts and will be provided with the bulk shard operation requests. The bulk shard operations api will fail if the provided history uuid is unequal to the actual history uuid.
No longer record the leader history uuid in shard follow task params, but rather use the leader history UUIDs directly from follower index's custom metadata. The resume follow api will remain to fail if leader index shard history UUIDs are missing.
Closes#33956
Since #34099, the FollowingEngine will skip an operation which was
already processed before. With that change, it should be okay to unmute
testFollowIndexAndCloseNode.
This arose when two commits were pushed at roughly the same time, both
of which compiled successfully against master, but not when taken
together. This commit fixes a reference in one of the commits that was
changed in the other commit.
This commit modifies the CCR stats endpoint for indices to be
/{index}/_ccr/stats. This makes this endpoint consistent with other
index-centric endpoints like indices stats.
The unfollow API changes a follower index into a regular index, so that it will accept write requests from clients.
For the unfollow api to work the index follow needs to be stopped and the index needs to be closed.
Closes#33931
This change introduces the indexing optimization using sequence numbers
in the FollowingEngine. This optimization uses the max_seq_no_updates
which is tracked on the primary of the leader and replicated to replicas
and followers.
Relates #33656
Prior to following an index in the follow API, check whether current
user has sufficient privileges in the leader cluster to read and
monitor the leader index.
Also check this in the create and follow API prior to creating the
follow index.
Also introduced READ_CCR cluster privilege that include the minimal
cluster level actions that are required for ccr in the leader cluster.
So a user can follow indices in a cluster, but not use the ccr admin APIs.
Closes#33553
Co-authored-by: Jason Tedor <jason@tedor.me>
The following changes were made:
* Added ElasticsearchSecurityException. For in the case the current user has insufficient privileges while an index is being followed. Prior to following ccr checks whether the current user has sufficient privileges and if not the follow api fails with an error.
* Added Index block exception. If the leader index gets closed, this exception is returned.
* Added ClusterBlockException service unavailable. In case for example the leader cluster is without elected master.
* Removed IndexNotFoundException. If the leader / follower index has been deleted, ccr will need to stop the shard follow tasks with an error.
Closes#33954
* Renamed CCR APIs
Renamed:
* `/{index}/_ccr/create_and_follow` to `/{index}/_ccr/follow`
* `/{index}/_ccr/unfollow` to `/{index}/_ccr/pause_follow`
* `/{index}/_ccr/follow` to `/{index}/_ccr/resume_follow`
Relates to #33931
always use `IndicesOptions.strictExpand()` for indices options.
The follow index may be closed and we still want to get stats from
shard follow task and the whether the provided index name matches with
follow index name is checked when locating the task itself in the ccr
stats transport action.
This commit replicates the max_seq_no_of_updates on the leading index
to the primaries of the following index via ShardFollowNodeTask. The
max_seq_of_updates is then transmitted to the replicas of the follower
via replication requests (that's BulkShardOperationsRequest).
Relates #33656
We start tracking max seq_no_of_updates on the primary in #33842. This
commit replicates that value from a primary to its replicas in replication
requests or the translog phase of peer-recovery.
With this change, we guarantee that the value of max seq_no_of_updates
on a replica when any index/delete operation is performed at least the
max_seq_no_of_updates on the primary when that operation was executed.
Relates #33656
If numWrites is between 2 and 9, we will issue an invalid range because
the from_seq_no is negative. This commit makes sure that numWrites is at
least 10, and adds an explicit test to verify invalid request ranges.
This PR is the first step to use seq_no to optimize indexing operations.
The idea is to track the max seq_no of either update or delete ops on a
primary, and transfer this information to replicas, and replicas use it
to optimize indexing plan for index operations (with assigned seq_no).
The max_seq_no_of_updates on primary is initialized once when a primary
finishes its local recovery or peer recovery in relocation or being
promoted. After that, the max_seq_no_of_updates is only advanced internally
inside an engine when processing update or delete operations.
Relates #33656
Instead of having one constructor that accepts all arguments, all parameters
should be provided via setters. Only leader and follower index are required
arguments. This makes using this class in tests and transport client easier.
The following stats are being kept track of:
1) The total number of times that auto following a leader index succeed.
2) The total number of times that auto following a leader index failed.
3) The total number of times that fetching a remote cluster state failed.
4) The most recent 256 auto follow failures per auto leader index
(e.g. create_and_follow api call fails) or cluster alias
(e.g. fetching remote cluster state fails).
Each auto follow run now produces a result that is being used to update
the stats being kept track of in AutoFollowCoordinator.
Relates to #33007
* [CCR] Do not unnecessarily wrap fetch exception in a ElasticSearch exception and
properly map fetch_exception.exception field as object.
The extra caused by level is not necessary here:
```
"fetch_exceptions": [
{
"from_seq_no": 1,
"retries": 106,
"exception": {
"type": "exception",
"reason": "[index1] IndexNotFoundException[no such index]",
"caused_by": {
"type": "index_not_found_exception",
"reason": "no such index",
"index_uuid": "_na_",
"index": "index1"
}
}
}
],
```
When a leader index is created, it may not have a mapping yet.
Currently if you follow such an index the shard follow tasks fail with
NoSuchElementException, because they expect a single mapping.
This commit fixes that, by allowing that a leader index does not yet have
a mapping.
Rather than scheduling pings to the leader index when we are caught up
to the leader, this commit introduces long polling for changes. We will
fire off a request to the leader which if we are already caught up will
enter a poll on the leader side to listen for global checkpoint
changes. These polls will timeout after a default of one minute, but can
also be specified when creating the following task. We use these time
outs as a way to keep statistics up to date, to not exaggerate time
since last fetches, and to avoid pipes being broken.
When executing CCR REST tests it is going to be expected after global
checkpoint polling goes in that shard changes tasks can still be pending
at the end of the test. One way to deal with this is to set a low
timeout on these polls, but then that means we are not executing our
REST tests with our default production settings and instead would be
using an unrealistic low timeout. Alternatively, since we expect these
tasks to be there, we can not count them against the test. That is what
this commit does.
This commit moves these REST tests (possibly temporarily) to a
sub-project of ccr. We do this (again, possibly temporarily) to keep
them within the ccr sub-project yet there are changes within 6.x that
prevent these from being in the top-level project (the cluster formation
tasks are trying to install x-pack-ccr into the
integ-test-zip). Therefore, we isolate these for now until we can
understand why there are differences between 6.x and master.
When developing ccr it is not ideal if tests are in multiple modules.
Even the classes these tests test are in the core module, it is easier
if these tests are in ccr module in order to avoid running the test task
in core module. This results in running many non ccr tests.
This way when developing ccr we can run locally:
./gradlew x-pack:plugin:core:precommit x-pack:plugin:ccr:check
before pushing to PR branches and be confident that the PR build passes,
without running x-pack:plugin:core:check task.
and if so debug log it and otherwise rethrow.
This should fix a couple of test failures where during test teardown tests
failed due to uncaught exceptions being detected.
The follow index api checks if the recorded uuid in the follow index matches
with uuid of the leader index and fails otherwise. This validation will
prevent a follow index from following an incompatible leader index.
The create_and_follow api will automatically add this custom index metadata
when it creates the follow index.
Closes#31505
Martijn van Groningen
Alpar Torok
Jason Tedor
Jim Ferenczi
Luca Cavanna
Nhat Nguyen
Tim Brooks
Boaz Leskes
Albert Zaharovits
Armin Braun
Andrey Ershov
Ryan Ernst
Yannick Welsch
Adrien Grand
David Turner
Arthur Gavlyukovskiy
Nik Everett
Alexander Reelsen
David Kyle
Andrey Atapin