Do not remove flood block from indices on nodes undergoing replacement #78942
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This commit enhances `DiskThresholdMonitor` so that indices that have a flood-stage block will not have the block removed while they reside on a node that is part of a "REPLACE"-type node shutdown. This prevents a situation where a node is blocked due to disk usage, then during the replacement the block is removed while shards are relocating to the target node, indexing occurs, and then the target runs out of space due to the additional documents. Relates to elastic#70338 and elastic#76247
dakrone
added
v8.0.0
:Core/Infra/Node Lifecycle
|
Pinging @elastic/es-core-infra (Team:Core/Infra) |
henningandersen
approved these changes
Oct 12, 2021
| .collect(Collectors.toSet()); | ||
|
|
||
| // Generate a set of all the indices that exist on either the target or source of a node replacement | ||
| final Set<String> indicesOnReplaceSourceOrTarget = nodesIdsPartOfReplacement.stream() |
There was a problem hiding this comment.
I wonder if we should special handle the case where the source is empty (no assigned shards) and allow releasing the block then? That should be safe and could shorten the time until releasing the block in case there are multiple orchestration steps involved.
There was a problem hiding this comment.
I think we should leave it as-is for now, since we don't expect to be in the COMPLETE phase for a replacement for more than ~10ish seconds anyway, and if we determine that it is too long in the future, we can revisit this in subsequent work.
...er/src/test/java/org/elasticsearch/cluster/routing/allocation/DiskThresholdMonitorTests.java
Outdated
Show resolved
Hide resolved
...er/src/test/java/org/elasticsearch/cluster/routing/allocation/DiskThresholdMonitorTests.java
Outdated
Show resolved
Hide resolved
💔 Backport failed
You can use sqren/backport to manually backport by running |
dakrone
added a commit
to dakrone/elasticsearch
that referenced
this pull request
Oct 12, 2021
elastic#78942) This commit enhances `DiskThresholdMonitor` so that indices that have a flood-stage block will not have the block removed while they reside on a node that is part of a "REPLACE"-type node shutdown. This prevents a situation where a node is blocked due to disk usage, then during the replacement the block is removed while shards are relocating to the target node, indexing occurs, and then the target runs out of space due to the additional documents. Relates to elastic#70338 and elastic#76247 # Conflicts: # server/src/test/java/org/elasticsearch/cluster/routing/allocation/DiskThresholdMonitorTests.java
elasticsearchmachine
pushed a commit
that referenced
this pull request
Oct 12, 2021
#78942) (#79008) This commit enhances `DiskThresholdMonitor` so that indices that have a flood-stage block will not have the block removed while they reside on a node that is part of a "REPLACE"-type node shutdown. This prevents a situation where a node is blocked due to disk usage, then during the replacement the block is removed while shards are relocating to the target node, indexing occurs, and then the target runs out of space due to the additional documents. Relates to #70338 and #76247 # Conflicts: # server/src/test/java/org/elasticsearch/cluster/routing/allocation/DiskThresholdMonitorTests.java
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This commit enhances
DiskThresholdMonitorso that indices that have a flood-stage block will nothave the block removed while they reside on a node that is part of a "REPLACE"-type node shutdown.
This prevents a situation where a node is blocked due to disk usage, then during the replacement the
block is removed while shards are relocating to the target node, indexing occurs, and then the
target runs out of space due to the additional documents.
Relates to #70338 and #76247