The unassinged list is used to make allocation decisions but is currently
modified during allocation runs which causes primaries to be throttled
during allocation. If this happens newly allocated indices can be stalled
for a long time turning a cluster into a RED state if concurrent relocations
and / or recoveries are happening.
Closes#3610
Whether we remove the top-level folder from the archive depends now on the zip itself and not on where it was downloaded from. That makes it work installing local files too.
Closes#3582
Restrict the size of the input length to a reasonable size otherwise very
long strings can cause StackOverflowExceptions deep down in lucene land.
Yet, this is simply a saftly limit set to `50` UTF-16 codepoints by default.
This limit is only present at index time and not at query time. If prefix
completions > 50 UTF-16 codepoints are expected / desired this limit should be raised.
Critical string sizes are beyone the 1k UTF-16 Codepoints limit.
Closes#3596
Added more debug logging to TransportShardReplicationOperationAction
*Changed* node naming in tests from GUIDs to node# based naming as they are much easier to read
This test will make random allocation decision on a growing and
shrinking cluster leading to a random distribution of the shards.
After a certain amount of iterations the test allows allocation unless
the same shard is already allocated on a node and balances the cluster
to gain optimal balance.
The goal is to throw out suggestions that only meet the cutoff in some
shards. This will happen if your input phrase is only contained in a
few shards. If your shards are unbanced this rechecking can throw out
good suggestions.
Closes#3547.
If a path to the test classes contains spaces Hunspell tests failed due
to URL encoded paths. This happens on CI builds if you give the build
a name containing a space. This is fixed by first converting to a URI
and created a File object from the URI directly.
The fact that Maven and Eclipse share the same build directories can trigger
race conditions when both are trying to build at the same time, eg. if you run
`mvn clean test` while Eclipse is up and running: Eclipse will notice that some
class files are missing and start compiling in parallel with Maven.
In a special case if allocations from the "heaviest" to the "lighter" nodes is not possible
due to some allocation decider restrictions like zone awareness. if one zone has for instance
less nodes than another zone so one zone is horribly overloaded from a balanced perspective but we
can't move to the "lighter" shards since otherwise the zone would go over capacity.
Closes#3580
When recovery of a shard fails on a node, the node sends notification to the master with information about the failure. During the period between the shard failure and the time when notification about the failure reaches the master, any changes in shard allocations can trigger the node with the failed shard to allocate this shard again. This allocation (especially if successful) creates a ripple effect of the shard going through failure/started states in order to match the delayed states processed by master. Under certain condition, a node involved in this process might generate warning messages: "marked shard as started, but shard has not been created, mark shard as failed".
This fix makes sure that nodes keep track of failed shard allocations and will not try to allocate such shards repeatedly while waiting for the failure notification to be processed by master.
Clinton Gormley
Martijn van Groningen
Simon Willnauer
Luca Cavanna
Shay Banon
Boaz Leskes
Nik Everett
Adrien Grand
Igor Motov