The processors in the 1.7.x versions and before were designed to commit suicide if they encountered serious resource problems like NFS and database failures. Now in the HBase world, the 1.7.x proclivity to quitting at the first sign of trouble is not working out well. Back port the suicide protections that the 1.8 series has to the 1.7.X codebase.
Assignee: nobody → lars
Severity: normal → critical
Target Milestone: --- → 1.7.5
"it gets better." this is now in staging... I need to find a way to simulate hbase trouble in order to verify that this is working. In my test environment, I just set the timeout to an unrealistically low level and verified that it behaved as I expected.
this is tested in the 1.7.5 staging environment and it worked just fine.
Status: NEW → RESOLVED
Last Resolved: 8 years ago
Resolution: --- → FIXED
Component: Socorro → General
Product: Webtools → Socorro
You need to log in before you can comment on or make changes to this bug.