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
{{ message }}
This repository was archived by the owner on Oct 11, 2025. It is now read-only.
Hi,
we are seeing a number of cases where Pipelines seem to die because their tasks fail (and thus hit the max failuires in the queue). The main offender seems to be calls to ModulesService#getVersionHostname, which are done for each task instead of being set pipeline-wide.
Seems like a similar change (i.e. calling once per pipeline) has been implemented in mapreduce2 already.
Hi,
we are seeing a number of cases where Pipelines seem to die because their tasks fail (and thus hit the max failuires in the queue). The main offender seems to be calls to ModulesService#getVersionHostname, which are done for each task instead of being set pipeline-wide.
Seems like a similar change (i.e. calling once per pipeline) has been implemented in mapreduce2 already.
We are using the Java version.