[
https://issues.apache.org/jira/browse/HBASE-1023?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12650322#action_12650322
]
stack commented on HBASE-1023:
------------------------------
Global flusher should not have hardcoded values but rather should be
percentages of configured heap. Also, current low-water mark seems austere at
half the the high-water mark; could be a higher (Keep flushing till usage at
512 goes to 256M). Would suggest that as part of this patch, replace the high
and low MB with high and low percentages of heap.
> Check global flusher
> --------------------
>
> Key: HBASE-1023
> URL: https://issues.apache.org/jira/browse/HBASE-1023
> Project: Hadoop HBase
> Issue Type: Improvement
> Reporter: stack
> Fix For: 0.19.0
>
>
> It doesn't seem to be responsive:
> {code}
> 2008-11-22 16:48:53,135 INFO
> org.apache.hadoop.hbase.regionserver.MemcacheFlusher: Flushing
> streamitems,(�,1227400186823 because over global limit of 536870912
> 2008-11-22 16:48:53,135 DEBUG org.apache.hadoop.hbase.regionserver.HRegion:
> Started memcache flush for region streamitems,(�,1227400186823. Current
> region memcache size 52.8m
> 2008-11-22 16:48:55,385 DEBUG org.apache.hadoop.hbase.regionserver.HStore:
> Added /hbase/streamitems/813580134/content/mapfiles/5603572683448576472 with
> 86534 entries, sequence id 12187148, data size 52.8m, file size 54.2m to
> streamitems,(�,1227400186823
> 2008-11-22 16:48:55,385 DEBUG org.apache.hadoop.hbase.regionserver.HRegion:
> Finished memcache flush for region streamitems,(�,1227400186823 in 2250ms,
> sequence id=12187148, compaction requested=true
> 2008-11-22 16:48:55,385 INFO
> org.apache.hadoop.hbase.regionserver.MemcacheFlusher: Flushing
> streamitems,8�!,1227399924602 because over global limit of 536870912
> 2008-11-22 16:48:55,385 DEBUG org.apache.hadoop.hbase.regionserver.HRegion:
> Started memcache flush for region streamitems,8�!,1227399924602. Current
> region memcache size 52.6m
> 2008-11-22 16:48:58,588 DEBUG org.apache.hadoop.hbase.regionserver.HStore:
> Added /hbase/streamitems/1172538122/content/mapfiles/7009988811547233403 with
> 99947 entries, sequence id 12187149, data size 52.6m, file size 54.2m to
> streamitems,8�!,1227399924602
> 2008-11-22 16:48:58,588 DEBUG org.apache.hadoop.hbase.regionserver.HRegion:
> Finished memcache flush for region streamitems,8�!,1227399924602 in 3203ms,
> sequence id=12187149, compaction requested=true
> 2008-11-22 16:48:58,588 INFO
> org.apache.hadoop.hbase.regionserver.MemcacheFlusher: Flushing
> streamitems,Vd�,1227400051975 because over global limit of 536870912
> 2008-11-22 16:48:58,588 DEBUG org.apache.hadoop.hbase.regionserver.HRegion:
> Started memcache flush for region streamitems,Vd�,1227400051975. Current
> region memcache size 52.0m
> 2008-11-22 16:49:03,572 DEBUG org.apache.hadoop.hbase.regionserver.HStore:
> Added /hbase/streamitems/726818092/content/mapfiles/403342470777430233 with
> 118178 entries, sequence id 12202257, data size 52.0m, file size 53.9m to
> streamitems,Vd�,1227400051975
> 2008-11-22 16:49:03,603 DEBUG org.apache.hadoop.hbase.regionserver.HRegion:
> Finished memcache flush for region streamitems,Vd�,1227400051975 in 5015ms,
> sequence id=12202257, compaction requested=false
> 2008-11-22 16:49:03,603 INFO
> org.apache.hadoop.hbase.regionserver.MemcacheFlusher: Flushing
> streamitems,��+,1227400373676 because over global limit of 536870912
> 2008-11-22 16:49:03,603 DEBUG org.apache.hadoop.hbase.regionserver.HRegion:
> Started memcache flush for region streamitems,��+,1227400373676. Current
> region memcache size 44.8m
> 2008-11-22 16:49:06,464 DEBUG org.apache.hadoop.hbase.regionserver.HStore:
> Added /hbase/streamitems/1145535262/content/mapfiles/6954031123453241243 with
> 103669 entries, sequence id 12229234, data size 44.8m, file size 46.5m to
> streamitems,��+,1227400373676
> 2008-11-22 16:49:06,464 DEBUG org.apache.hadoop.hbase.regionserver.HRegion:
> Finished memcache flush for region streamitems,��+,1227400373676 in 2861ms,
> sequence id=12229234, compaction requested=false
> 2008-11-22 16:49:06,464 INFO
> org.apache.hadoop.hbase.regionserver.MemcacheFlusher: Flushing
> streamitems,i�k,1227399955582 because over global limit of 536870912
> 2008-11-22 16:49:06,464 DEBUG org.apache.hadoop.hbase.regionserver.HRegion:
> Started memcache flush for region streamitems,i�k,1227399955582. Current
> region memcache size 40.7m
> 2008-11-22 16:49:08,239 DEBUG org.apache.hadoop.hbase.regionserver.HStore:
> Added /hbase/streamitems/192477361/content/mapfiles/1760086866679202983 with
> 86915 entries, sequence id 12230327, data size 40.7m, file size 42.1m to
> streamitems,i�k,1227399955582
> 2008-11-22 16:49:08,239 DEBUG org.apache.hadoop.hbase.regionserver.HRegion:
> Finished memcache flush for region streamitems,i�k,1227399955582 in 1775ms,
> sequence id=12230327, compaction requested=true
> 2008-11-22 16:49:08,239 INFO
> org.apache.hadoop.hbase.regionserver.MemcacheFlusher: Flushing
> streamitems,x�,1227400725833 because over global limit of 536870912
> 2008-11-22 16:49:08,239 DEBUG org.apache.hadoop.hbase.regionserver.HRegion:
> Started memcache flush for region streamitems,x�,1227400725833. Current
> region memcache size 38.5m
> 2008-11-22 16:49:09,569 DEBUG org.apache.hadoop.hbase.regionserver.HStore:
> Added /hbase/streamitems/675890649/content/mapfiles/3120613868583981443 with
> 67978 entries, sequence id 12230328, data size 38.5m, file size 39.5m to
> streamitems,x�,1227400725833
> 2008-11-22 16:49:09,570 DEBUG org.apache.hadoop.hbase.regionserver.HRegion:
> Finished memcache flush for region streamitems,x�,1227400725833 in 1331ms,
> sequence id=12230328, compaction requested=false
> 2008-11-22 16:49:24,241 INFO org.apache.hadoop.hbase.regionserver.LogRoller:
> Rolling hlog. Number of entries: 100002
> {code}
> Why all the flushes? Why is global size same though we've dumped some
> flushes?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
