missing minor compaction files under heavy namenode load
--------------------------------------------------------
Key: ACCUMULO-65
URL: https://issues.apache.org/jira/browse/ACCUMULO-65
Project: Accumulo
Issue Type: Bug
Components: tserver
Affects Versions: 1.3.5, 1.4.0
Environment: dynamic ingest on a large cluster, hadoop cdh3beta2
Reporter: Eric Newton
Assignee: Eric Newton
Monitor was showing lots of errors for missing files. Analysis of any single
file shows that the tablet was not moved, nor assigned to multiple servers.
All the errors are for files that were minor compacted with many namenode
operations failing/retried. The files were not deleted by the accumulo garbage
collector. Checking the name node logs, there is no mention of the file being
created, but there is a mention of the final rename of the file failing.
Possible HDFS issue: file open and write succeeds, close succeeds, the file is
then re-opened, and checked; yet the file is not created.
The return code of the rename to bring the file online was not checked.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira