Setting `discovery.initial_state_timeout: 0s` to make `discovery.zen.minimum_master_nodes: N`
work reliably can cause issues in clusters that rely on state recovery once the cluster is available.
This change makes the use or `discovery.zen.minimum_master_nodes` optional for clusters where this behavior is desirable.
When installing the Windows service, certain settings like the minimum
heap, maximum heap and thread stack size setting must be set. While
there is an error message making mention of this fact, the error message
is not explicit exactly what setting needs to be set. This commit makes
these settings explicit.
Relates #21200
Before publishing a cluster state the master connects to the nodes that are added in the cluster state. When publishing fails, however, it does not disconnect from these nodes, leaving NodeConnectionsService out of sync with the currently applied cluster state.
The assertion assertMaster checks if all nodes have each other in the cluster state and the correct master set.
It is usually called after a disruption has been healed and ensureStableCluster been called. In presence of a low
publish timeout of 1s in this test class, publishing might not be fully done even after ensureStableCluster returns.
This commit adds an assertBusy to assertMaster so that the node has a bit more time to apply the cluster state from
the master, even if it's a bit slow.
Replication request may arrive at a replica before the replica's node has processed a required mapping update. In these cases the TransportReplicationAction will retry the request once a new cluster state arrives. Sadly that retry logic failed to call `ReplicationRequest#onRetry`, causing duplicates in the append only use case.
This commit fixes this and also the test which missed the check. I also added an assertion which would have helped finding the source of the duplicates.
This was discovered by https://elasticsearch-ci.elastic.co/job/elastic+elasticsearch+master+multijob-unix-compatibility/os=opensuse/174/
Relates #20211
Java 9's exception message when lists have an out of bounds index
is much better than java 8 but the painless code asserted on the
java 8 message. Now it'll accept either.
I'm tempted to weaken the assertion but I like asserting that the
message is readable.
Adds support for indexing into lists and arrays with negative
indexes meaning "counting from the back". So for if
`x = ["cat", "dog", "chicken"]` then `x[-1] == "chicken"`.
This adds an extra branch to every array and list access but
some performance testing makes it look like the branch predictor
successfully predicts the branch every time so there isn't a
in execution time for this feature when the index is positive.
When the index is negative performance testing showed the runtime
is the same as writing `x[x.length - 1]`, again, presumably thanks
to the branch predictor.
Those performance metrics were calculated for lists and arrays but
`def`s get roughly the same treatment though instead of inlining
the test they need to make a invoke dynamic so we don't screw up
maps.
Closes#20870
Vagrant tests use a static list of dependencies to upgrade from
and we weren't including 5.0.0 deps in that list. Also when the
list was incorrect we weren't sorting the "current" list so it
was difficult to read.
Also adds 2.4.1 to the list but *doesn't* add 5.0.0 because we
still can't resolve it yet. We still only print an error when
the list is wrong but don't abort the build. We'll abort the build
once we've fixed resolution for 5.0.0 and we can re-add it.
The `_cat/nodes` API might not be available in all clusters for instance
if they have authorization enabled. This change falls back to the previously
used method of using the '/' endpoint to fetch the nodes version, this is best
effort and will emit a warning.
We are upgrading from out of date versions in our tests right now and we
can't fix that because the current versions to upgrade from aren't in
maven central. We'll resolve the resolution issue soon, but for now
let's get the build green.
Lucene 6.3 is expected to be released in the next weeks so it'd be good to give
it some integration testing. I had to upgrade randomized-testing too so that
both Lucene and Elasticsearch are on the same version.
Today we only use a single node to send requests to when we run REST tests.
In some cases we have more than one node (ie. in the BWC case) where we should
send requests to all nodes in a round-robin fashion. This change passes all
available node endpoints to the rest test.
Additionally, this change adds the setting of `discovery.zen.minimum_master_nodes`
to the cluster formation forcing the nodes to wait for all other nodes until the cluster
is formed. This allows for a more realistic master election and allows all master eligable
nodes to become master while before always the first node in the cluster became the master.
This also adds logging to each test run to log the master nodes version and the minimum node
version in the cluster to help debugging BWC test failures.
This fixes our cluster formation task to run REST tests against a mixed version cluster.
Yet, due to some limitations in our test framework `indices.rollover` tests are currently
disabled for the BWC case since they select the current master as the merge node which
happens to be a BWC node and we can't relocate all shards to it since the primaries are on
a higher version node. This will be fixed in a followup.
Closes#21142
Note: This has been cherry-picked from 5.0 and fixes several rest tests
as well as a BWC break in `OsStats.java`
The network disruption type "network delay" continues delaying existing requests even after the disruption has been cleared. This commit ensures that the requests get to execute right after the delay rule is cleared.
This test failed when the node that was shutting down was not yet removed from the cluster state on the master.
The cluster allocation explain API will not see any unassigned shards until the node shutting down is removed from the
cluster state.
Previously, if a node left the cluster (for example, due to a long GC),
during a snapshot, the master node would mark the snapshot as failed, but
the node itself could continue snapshotting the data on its shards to the
repository. If the node rejoins the cluster, the master may assign it to
hold the replica shard (where it held the primary before getting kicked off
the cluster). The initialization of the replica shard would repeatedly fail
with a ShardLockObtainFailedException until the snapshot thread finally
finishes and relinquishes the lock on the Store.
This commit resolves the situation by ensuring that when a shard is removed
from a node (such as when a node rejoins the cluster and realizes it no longer
holds the active shard copy), any snapshotting of the removed shards is aborted.
In the scenario above, when the node rejoins the cluster, it will see in the cluster
state that the node no longer holds the primary shard, so IndicesClusterStateService
will remove the shard, thereby causing any snapshots of that shard to be aborted.
Closes#20876
The cluster state on a node is updated either
- by incoming cluster states that are received from the active master or
- by the node itself when it notices that the master has gone.
In the second case, the node adds the NO_MASTER_BLOCK and removes the current master as active master from its cluster state. In one particular case, it would also update the list of nodes, removing the master node that just failed. In the future, we want a clear separation between actions that can be executed by a master publishing a cluster state and a node locally updating its cluster state when no active master is around.
This commit fixes responses to HEAD requests so that the value of the
Content-Length is correct per the HTTP spec. Namely, the value of this
header should be equal to the Content-Length if the request were not a
HEAD request.
This commit also fixes a memory leak on HEAD requests to the main action
that arose from the bytes on a builder not being released due to them
being dropped on the floor to ensure that the response to the main
action did not have a body.
Relates #21123
Yannick Welsch
Lee Hinman
Stanislav Mamontov
Simon Willnauer
Jason Tedor
Boaz Leskes
Nik Everett
Igor Motov
Simon Willnauer
Ali Beyad
Adrien Grand
Areek Zillur
markharwood
Jun Ohtani
Colin Goodheart-Smithe
Jack Conradson
Christoph Büscher