[ https://issues.apache.org/jira/browse/HIVE-2866?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Phabricator updated HIVE-2866: ------------------------------ Attachment: HIVE-2866.D2277.1.patch kevinwilfong requested code review of "HIVE-2866 [jira] Cache local map reduce job errors for additional logging". Reviewers: JIRA https://issues.apache.org/jira/browse/HIVE-2866 The error stream from subprocesses where local map reduce jobs are run are now passed through a Caching PrintStream. If the job fails, the output written to this print stream are stored in the SessionState in a map from stage ID to a list of strings, which represents the lines of output written to the error stream. This way, these can be logged along with the failed query in some central location for easier debugging. Using the CachingPrintStream we are storing errors that occur locally in Hive, but because local map reduce jobs are run in a separate JVM we are not storing that occur for these. We can use this same construct to store errors written to the subprocesses error stream. This way, when we log failed queries, these will give us a decent idea of why those queries failed. See related issues: https://issues.apache.org/jira/browse/HIVE-2832 https://issues.apache.org/jira/browse/HIVE-2858 TEST PLAN EMPTY REVISION DETAIL https://reviews.facebook.net/D2277 AFFECTED FILES data/scripts/cat_error.py common/src/java/org/apache/hadoop/hive/common/io/CachingPrintStream.java ql/src/test/results/clientnegative/local_mapred_error_cache.q.out ql/src/test/org/apache/hadoop/hive/ql/hooks/VerifySessionStateLocalErrorsHook.java ql/src/test/queries/clientnegative/local_mapred_error_cache.q ql/src/java/org/apache/hadoop/hive/ql/session/SessionState.java ql/src/java/org/apache/hadoop/hive/ql/exec/MapRedTask.java ql/src/java/org/apache/hadoop/hive/ql/exec/MapredLocalTask.java ql/src/java/org/apache/hadoop/hive/ql/Driver.java MANAGE HERALD DIFFERENTIAL RULES https://reviews.facebook.net/herald/view/differential/ WHY DID I GET THIS EMAIL? https://reviews.facebook.net/herald/transcript/5061/ Tip: use the X-Herald-Rules header to filter Herald messages in your client. > Cache local map reduce job errors for additional logging > -------------------------------------------------------- > > Key: HIVE-2866 > URL: https://issues.apache.org/jira/browse/HIVE-2866 > Project: Hive > Issue Type: Improvement > Components: Logging > Reporter: Kevin Wilfong > Assignee: Kevin Wilfong > Attachments: HIVE-2866.D2277.1.patch > > > Using the CachingPrintStream we are storing errors that occur locally in > Hive, but because local map reduce jobs are run in a separate JVM we are not > storing that occur for these. We can use this same construct to store errors > written to the subprocesses error stream. This way, when we log failed > queries, these will give us a decent idea of why those queries failed. > See related issues: > https://issues.apache.org/jira/browse/HIVE-2832 > https://issues.apache.org/jira/browse/HIVE-2858 -- 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