Getting index statistics could contain failed shards #2210
Merged
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
During rotation and retention we requested all index statistics multiple times, which, in a overloaded cluster, could lead to shard failures due to timeouts. This failure wasn't logged and could lead to using the wrong (older) index to base rotation decisions on, effectively rotating indices too early. This change makes Graylog use a more lightweight API to determine all index names including their aliases, reducing the usage of the expensive Index Statistics to the indices page only. The current rotation and retention strategies do not need to know all index statistics which require to touch every single shard in the cluster. fixes #2194
| } catch (NumberFormatException ex) { | ||
| LOG.debug("Couldn't extract index number from index name " + indexName, ex); | ||
| LOG.warn("Couldn't extract index number from index name " + indexName, ex); | ||
| } | ||
| } | ||
|
|
There was a problem hiding this comment.
What about replacing this with a streams call like this:
final Optional<Integer> highestIndexNumber = indexNames.stream()
.filter(indexName -> !this.isGraylogDeflectorIndex(indexName))
.map(Deflector::extractIndexNumber)
.max(Integer::max);
Makes it easier to grasp what is done (at least for me).
There was a problem hiding this comment.
In theory I agree, but in this case we'd need to refactor extractIndexNumber as well and its callers, because it throws an exception, and it already felt I changed a lot of code already :(
How about making those changes on master?
|
LGTM 👍 |
bernd
pushed a commit
that referenced
this pull request
May 11, 2016
During rotation and retention we requested all index statistics multiple times, which, in a overloaded cluster, could lead to shard failures due to timeouts. This failure wasn't logged and could lead to using the wrong (older) index to base rotation decisions on, effectively rotating indices too early. This change makes Graylog use a more lightweight API to determine all index names including their aliases, reducing the usage of the expensive Index Statistics to the indices page only. The current rotation and retention strategies do not need to know all index statistics which require to touch every single shard in the cluster. Fixes #2194 (cherry picked from commit bc6042c)
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.
During rotation and retention we requested all index statistics multiple times, which, in a overloaded cluster, could lead to shard failures due to timeouts.
This failure wasn't logged and could lead to using the wrong (older) index to base rotation decisions on, effectively rotating indices too early.
This change makes Graylog use a more lightweight API to determine all index names including their aliases, reducing the usage of the expensive Index Statistics to the indices page only.
The current rotation and retention strategies do not need to know all index statistics which require to touch every single shard in the cluster.
fixes #2194