You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Recently I've seen requests trigger a failure, where MVEL gets into an infinite loop trying to report the exception.
This failure causes service to degrade across the entire cluster, causing timeouts on requests handled by all nodes. These may be just for indexes with a shard on the affected machine, but it feels bigger than that.
Based on @nz 's report, I've submitted a patch to MVEL to solve this issue (http://jira.codehaus.org/browse/MVEL-292 ) . It is limited in scope to a single line script that start with a new line and trigger an error while executing/parsing (this example, a bad field data access).
Recently I've seen requests trigger a failure, where MVEL gets into an infinite loop trying to report the exception.
This failure causes service to degrade across the entire cluster, causing timeouts on requests handled by all nodes. These may be just for indexes with a shard on the affected machine, but it feels bigger than that.
Shared some logs and a repro with @kimchy.
The text was updated successfully, but these errors were encountered: