[
https://issues.apache.org/jira/browse/HBASE-2457?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12857683#action_12857683
]
stack commented on HBASE-2457:
------------------------------
Here is the code:
{code}
if (!majorcompaction && !references) {
// Here we select files for incremental compaction.
// The rule is: if the largest(oldest) one is more than twice the
// size of the second, skip the largest, and continue to next...,
// until we meet the compactionThreshold limit.
for (point = 0; point < countOfFiles - 1; point++) {
if ((fileSizes[point] < fileSizes[point + 1] * 2) &&
(countOfFiles - point) <= maxFilesToCompact) {
break;
}
skipped += fileSizes[point];
}
filesToCompact = new ArrayList<StoreFile>(filesToCompact.subList(point,
countOfFiles));
...
{code}
Todd did a listing of the Store:
{code}
-rw-r--r-- 3 todd supergroup 134676288 2010-04-15 20:58
/hbase/test1/1736416594/actions/7876304170849844721
-rw-r--r-- 3 todd supergroup 61309324 2010-04-15 20:59
/hbase/test1/1736416594/actions/3289606064411356106
-rw-r--r-- 3 todd supergroup 24381446 2010-04-15 21:00
/hbase/test1/1736416594/actions/4995952634622872563
-rw-r--r-- 3 todd supergroup 1194376 2010-04-15 21:01
/hbase/test1/1736416594/actions/3688488928995595533
-rw-r--r-- 3 todd supergroup 2712677 2010-04-15 21:01
/hbase/test1/1736416594/actions/5274234463019618354
-rw-r--r-- 3 todd supergroup 532824 2010-04-15 21:01
/hbase/test1/1736416594/actions/5321716733066884905
-rw-r--r-- 3 todd supergroup 12103493 2010-04-15 21:01
/hbase/test1/1736416594/actions/6876448782185582993
-rw-r--r-- 3 todd supergroup 5855317 2010-04-15 21:01
/hbase/test1/1736416594/actions/7381870009831588255
{code}
If you trace, the olders is > 2 * the next oldest, and so on. Some of the
times are same so not sure how that plays out.... (the above is not strictly
ordered ... those of same time may not be proper chronological order).
> RS gets stuck compacting region ad infinitum
> --------------------------------------------
>
> Key: HBASE-2457
> URL: https://issues.apache.org/jira/browse/HBASE-2457
> Project: Hadoop HBase
> Issue Type: Bug
> Affects Versions: 0.20.4
> Reporter: Todd Lipcon
> Priority: Critical
> Attachments: log.gz, stack
>
>
> Testing 0.20_pre_durabil...@934643, I ended up in a state where one region
> server got stuck compacting a single region over and over again forever. This
> was with a special config with very low flush threshold in order to stress
> test flush/compact code.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
