When a cluster is restarted, jetty is not functioning for a small percentage of datanodes, requiring restart of those datanodes. This is caused by JETTY-1316.
We've tried overriding isRunning() and retrying on super.isRunning() returning false, as the reporter of JETTY-1316 mentioned in the description. It looks like the code was actually exercised (i.e. the issue was caused by this jetty bug) and the acceptor was working fine after retry.
Since we will probably move to a later version of jetty after branch-3 is cut, we can put this workaround in branch-2 only.
|Transition||Time In Source Status||Execution Times||Last Executer||Last Execution Date|
|6d 23h 16m||1||Kihwal Lee||15/May/14 21:38|
|91d 8h||1||Karthik Kambatla (Inactive)||15/Aug/14 05:39|
|Status||Resolved [ 5 ]||Closed [ 6 ]|
|Status||Open [ 1 ]||Resolved [ 5 ]|
|Hadoop Flags||Reviewed [ 10343 ]|
|Fix Version/s||0.23.11 [ 12324665 ]|
|Fix Version/s||2.5.0 [ 12326263 ]|
|Resolution||Fixed [ 1 ]|
|Summary||Workaound for jetty6 acceptor startup issue||Workaround for jetty6 acceptor startup issue|
|Assignee||Kihwal Lee [ kihwal ]|