You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hbase.apache.org by "Billy Pearson (JIRA)" <ji...@apache.org> on 2008/08/10 17:09:44 UTC

[jira] Created: (HBASE-812) Compaction needs little better skip algo

Compaction needs little better skip algo
----------------------------------------

                 Key: HBASE-812
                 URL: https://issues.apache.org/jira/browse/HBASE-812
             Project: Hadoop HBase
          Issue Type: Bug
          Components: regionserver
    Affects Versions: 0.2.0, 0.3.0
            Reporter: Billy Pearson
            Priority: Trivial
             Fix For: 0.2.1, 0.3.0


Looking at this section of one of my compaction's we have 3 files to compact the new algo is working great in my test but I see this below often we are skipping 2 out of the 3 files and compacting 1 file. 1 file is kind of a wast might as well just copy the file my suggestion is if there is only 1 file left after the new algo skips then just go on to the next column and skip the last file also. This will help improve compaction times a little more. 

{code}
2008-08-10 10:00:45,310 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 1339600874/size: 4.6m, skipped 2, 4851776
2008-08-10 10:00:45,438 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 1 files into /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891
2008-08-10 10:00:46,838 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891 to /hbase/webdata/1339600874/size/mapfiles/7539342470259528578
2008-08-10 10:00:47,166 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 1339600874/size store size is 4.6m
{code}

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Updated: (HBASE-812) Compaction needs little better skip algo

Posted by "Billy Pearson (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HBASE-812?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Billy Pearson updated HBASE-812:
--------------------------------

    Status: Patch Available  (was: Open)

I made a new patch to add a new line on the debug logging its kind of hard to catch those 3 words on the end of the line and know whats going on and fixed a space missed on the old log line you edited. 
Fill free to change the text but I would like to see a new line saying we skipped the compaction of only 1 file.

but patch works great does just what was needed!

 

> Compaction needs little better skip algo
> ----------------------------------------
>
>                 Key: HBASE-812
>                 URL: https://issues.apache.org/jira/browse/HBASE-812
>             Project: Hadoop HBase
>          Issue Type: Bug
>          Components: regionserver
>    Affects Versions: 0.2.0, 0.3.0
>            Reporter: Billy Pearson
>            Priority: Critical
>             Fix For: 0.2.1, 0.3.0
>
>         Attachments: 812.patch.txt, 812v2.patch
>
>
> Looking at this section of one of my compaction's we have 3 files to compact the new algo is working great in my test but I see this below often we are skipping 2 out of the 3 files and compacting 1 file. 1 file is kind of a wast might as well just copy the file my suggestion is if there is only 1 file left after the new algo skips then just go on to the next column and skip the last file also. This will help improve compaction times a little more. 
> {code}
> 2008-08-10 10:00:45,310 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 1339600874/size: 4.6m, skipped 2, 4851776
> 2008-08-10 10:00:45,438 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 1 files into /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891
> 2008-08-10 10:00:46,838 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891 to /hbase/webdata/1339600874/size/mapfiles/7539342470259528578
> 2008-08-10 10:00:47,166 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 1339600874/size store size is 4.6m
> {code}

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Commented: (HBASE-812) Compaction needs little better skip algo

Posted by "Billy Pearson (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HBASE-812?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12621608#action_12621608 ] 

Billy Pearson commented on HBASE-812:
-------------------------------------

I am running the 0.2.0 release and no its not compacting a reference. after a split we do not do the skip algo when we are compacting a reference its a force compaction of all the files. 
I have more examples thats just one of the columns but we are doing this a lot when I am running a import job.

But the skip algo is working great over all I see compaction's of about 1 min vs 5-7 mins per compaction while under insert load.
There is some in the below log I just found a group of compaction's and cut this out of the log I thank its after a startup but still you can see the skip algo picks 2 on 
some of them and about half it only has 1 left we just need to skip the columns that only have 1 mapfile after the algo does the selection of the files. then on the next 
compaction for the region it will have 2 mapfiles and do something when its compacting.



{code}
2008-08-10 23:56:13,563 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/200761489/anchor/mapfiles/4695022772964499808 to /hbase/webdata/200761489/anchor/mapfiles/6542744344240857591
2008-08-10 23:56:13,631 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 200761489/anchor store size is 72.8m
2008-08-10 23:56:13,637 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 200761489/stime: 5.1m, skipped 1, 4191303
2008-08-10 23:56:13,706 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 2 files into /hbase/webdata/compaction.dir/200761489/stime/mapfiles/6863349056798671266
2008-08-10 23:56:15,559 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/200761489/stime/mapfiles/6863349056798671266 to /hbase/webdata/200761489/stime/mapfiles/3093620455129104149
2008-08-10 23:56:15,639 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 200761489/stime store size is 5.1m
2008-08-10 23:56:15,647 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 200761489/in_rank: 62.5m, skipped 1, 55688666
2008-08-10 23:56:15,724 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 2 files into /hbase/webdata/compaction.dir/200761489/in_rank/mapfiles/1559371687332050999
2008-08-10 23:56:32,318 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/200761489/in_rank/mapfiles/1559371687332050999 to /hbase/webdata/200761489/in_rank/mapfiles/5872775000657698199
2008-08-10 23:56:32,400 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 200761489/in_rank store size is 61.1m
2008-08-10 23:56:32,407 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 200761489/size: 5.5m, skipped 1, 4465439
2008-08-10 23:56:32,472 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 2 files into /hbase/webdata/compaction.dir/200761489/size/mapfiles/3296022424629716093
2008-08-10 23:56:34,710 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/200761489/size/mapfiles/3296022424629716093 to /hbase/webdata/200761489/size/mapfiles/8583717577394542471
2008-08-10 23:56:34,782 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 200761489/size store size is 5.4m
2008-08-10 23:56:34,789 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 200761489/last_seen: 61.5m, skipped 1, 54824835
2008-08-10 23:56:34,872 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 2 files into /hbase/webdata/compaction.dir/200761489/last_seen/mapfiles/3943134197356605963
2008-08-10 23:56:50,981 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/200761489/last_seen/mapfiles/3943134197356605963 to /hbase/webdata/200761489/last_seen/mapfiles/307647539041731711
2008-08-10 23:56:51,219 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 200761489/last_seen store size is 60.1m
2008-08-10 23:56:51,242 INFO org.apache.hadoop.hbase.regionserver.HRegion: compaction completed on region webdata,de.popstars-new-angels.www%2Fpopstars-biographien.html%3Ahttp,1218422328154 in 52sec
2008-08-10 23:56:51,245 INFO org.apache.hadoop.hbase.regionserver.HRegion: starting compaction on region webdata,fr.asso.agirpourlenfant.www%2Findex.html%3Ahttp,1218422328154
2008-08-10 23:56:51,252 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 338914710/anchor: 81.6m, skipped 2, 85047362
2008-08-10 23:56:51,268 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 1 files into /hbase/webdata/compaction.dir/338914710/anchor/mapfiles/2546621217286575105
2008-08-10 23:56:52,143 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/338914710/anchor/mapfiles/2546621217286575105 to /hbase/webdata/338914710/anchor/mapfiles/3839219944106123077
2008-08-10 23:56:52,188 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 338914710/anchor store size is 80.4m
2008-08-10 23:56:52,195 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 338914710/stime: 5.6m, skipped 2, 5784755
2008-08-10 23:56:52,210 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 1 files into /hbase/webdata/compaction.dir/338914710/stime/mapfiles/1286630914618726188
2008-08-10 23:56:52,414 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/338914710/stime/mapfiles/1286630914618726188 to /hbase/webdata/338914710/stime/mapfiles/5121005992036000876
2008-08-10 23:56:52,456 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 338914710/stime store size is 5.5m
2008-08-10 23:56:52,462 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 338914710/in_rank: 71.3m, skipped 2, 74283017
2008-08-10 23:56:52,486 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 1 files into /hbase/webdata/compaction.dir/338914710/in_rank/mapfiles/1957109104676001133
2008-08-10 23:56:53,524 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/338914710/in_rank/mapfiles/1957109104676001133 to /hbase/webdata/338914710/in_rank/mapfiles/9145853810894409994
2008-08-10 23:56:53,569 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 338914710/in_rank store size is 69.8m
2008-08-10 23:56:53,576 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 338914710/size: 5.9m, skipped 2, 6167045
2008-08-10 23:56:53,589 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 1 files into /hbase/webdata/compaction.dir/338914710/size/mapfiles/5911079568309110370
2008-08-10 23:56:53,809 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/338914710/size/mapfiles/5911079568309110370 to /hbase/webdata/338914710/size/mapfiles/5518032354875458699
2008-08-10 23:56:53,853 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 338914710/size store size is 5.9m
2008-08-10 23:56:53,860 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 338914710/last_seen: 70.6m, skipped 2, 73521266
2008-08-10 23:56:53,902 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 1 files into /hbase/webdata/compaction.dir/338914710/last_seen/mapfiles/4617217208970844256
2008-08-10 23:56:54,898 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/338914710/last_seen/mapfiles/4617217208970844256 to /hbase/webdata/338914710/last_seen/mapfiles/6045607152317204195
2008-08-10 23:56:54,943 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 338914710/last_seen store size is 69.1m
2008-08-10 23:56:54,948 INFO org.apache.hadoop.hbase.regionserver.HRegion: compaction completed on region webdata,fr.asso.agirpourlenfant.www%2Findex.html%3Ahttp,1218422328154 in 3sec
2008-08-10 23:56:54,951 INFO org.apache.hadoop.hbase.regionserver.HRegion: starting compaction on region webdata,org.jessejacksonjr.www%2Fissues%2Fi0809001569.html%3Ahttp,1218423851154
2008-08-10 23:56:54,961 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 376160388/anchor: 70.0m, skipped 2, 72330348
2008-08-10 23:56:55,258 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 2 files into /hbase/webdata/compaction.dir/376160388/anchor/mapfiles/6390472557291621209
2008-08-10 23:56:56,953 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/376160388/anchor/mapfiles/6390472557291621209 to /hbase/webdata/376160388/anchor/mapfiles/6570235857466899342
2008-08-10 23:56:57,001 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 376160388/anchor store size is 69.0m
2008-08-10 23:56:57,008 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 376160388/stime: 4.5m, skipped 1, 4373520
2008-08-10 23:56:57,041 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 2 files into /hbase/webdata/compaction.dir/376160388/stime/mapfiles/8058281260768773414
2008-08-10 23:56:57,655 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/376160388/stime/mapfiles/8058281260768773414 to /hbase/webdata/376160388/stime/mapfiles/6655752511205630946
2008-08-10 23:56:57,710 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 376160388/stime store size is 4.4m
2008-08-10 23:56:57,719 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 376160388/in_rank: 60.7m, skipped 2, 62605727
2008-08-10 23:56:58,020 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 2 files into /hbase/webdata/compaction.dir/376160388/in_rank/mapfiles/6388173210597572693
2008-08-10 23:56:59,448 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/376160388/in_rank/mapfiles/6388173210597572693 to /hbase/webdata/376160388/in_rank/mapfiles/7217625771579221464
2008-08-10 23:56:59,498 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 376160388/in_rank store size is 59.4m
2008-08-10 23:56:59,504 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 376160388/size: 4.8m, skipped 1, 4651030
2008-08-10 23:56:59,912 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 2 files into /hbase/webdata/compaction.dir/376160388/size/mapfiles/6505866612377469653
2008-08-10 23:57:00,566 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/376160388/size/mapfiles/6505866612377469653 to /hbase/webdata/376160388/size/mapfiles/3951148926146160516
2008-08-10 23:57:00,612 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 376160388/size store size is 4.7m
2008-08-10 23:57:00,621 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 376160388/last_seen: 59.8m, skipped 2, 61665980
2008-08-10 23:57:00,814 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 2 files into /hbase/webdata/compaction.dir/376160388/last_seen/mapfiles/6278945227615596465
2008-08-10 23:57:02,202 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/376160388/last_seen/mapfiles/6278945227615596465 to /hbase/webdata/376160388/last_seen/mapfiles/4551300457309362976
2008-08-10 23:57:02,249 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 376160388/last_seen store size is 58.5m
2008-08-10 23:57:02,253 INFO org.apache.hadoop.hbase.regionserver.HRegion: compaction completed on region webdata,org.jessejacksonjr.www%2Fissues%2Fi0809001569.html%3Ahttp,1218423851154 in 7sec

{code}

> Compaction needs little better skip algo
> ----------------------------------------
>
>                 Key: HBASE-812
>                 URL: https://issues.apache.org/jira/browse/HBASE-812
>             Project: Hadoop HBase
>          Issue Type: Bug
>          Components: regionserver
>    Affects Versions: 0.2.0, 0.3.0
>            Reporter: Billy Pearson
>            Priority: Critical
>             Fix For: 0.2.1, 0.3.0
>
>
> Looking at this section of one of my compaction's we have 3 files to compact the new algo is working great in my test but I see this below often we are skipping 2 out of the 3 files and compacting 1 file. 1 file is kind of a wast might as well just copy the file my suggestion is if there is only 1 file left after the new algo skips then just go on to the next column and skip the last file also. This will help improve compaction times a little more. 
> {code}
> 2008-08-10 10:00:45,310 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 1339600874/size: 4.6m, skipped 2, 4851776
> 2008-08-10 10:00:45,438 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 1 files into /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891
> 2008-08-10 10:00:46,838 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891 to /hbase/webdata/1339600874/size/mapfiles/7539342470259528578
> 2008-08-10 10:00:47,166 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 1339600874/size store size is 4.6m
> {code}

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Commented: (HBASE-812) Compaction needs little better skip algo

Posted by "stack (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HBASE-812?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12621613#action_12621613 ] 

stack commented on HBASE-812:
-----------------------------

Thanks Billy.  Let me see if I can see whats broke (when you going to learn java?  So you can fix these things yourself (smile))

> Compaction needs little better skip algo
> ----------------------------------------
>
>                 Key: HBASE-812
>                 URL: https://issues.apache.org/jira/browse/HBASE-812
>             Project: Hadoop HBase
>          Issue Type: Bug
>          Components: regionserver
>    Affects Versions: 0.2.0, 0.3.0
>            Reporter: Billy Pearson
>            Priority: Critical
>             Fix For: 0.2.1, 0.3.0
>
>
> Looking at this section of one of my compaction's we have 3 files to compact the new algo is working great in my test but I see this below often we are skipping 2 out of the 3 files and compacting 1 file. 1 file is kind of a wast might as well just copy the file my suggestion is if there is only 1 file left after the new algo skips then just go on to the next column and skip the last file also. This will help improve compaction times a little more. 
> {code}
> 2008-08-10 10:00:45,310 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 1339600874/size: 4.6m, skipped 2, 4851776
> 2008-08-10 10:00:45,438 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 1 files into /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891
> 2008-08-10 10:00:46,838 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891 to /hbase/webdata/1339600874/size/mapfiles/7539342470259528578
> 2008-08-10 10:00:47,166 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 1339600874/size store size is 4.6m
> {code}

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Assigned: (HBASE-812) Compaction needs little better skip algo

Posted by "stack (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HBASE-812?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

stack reassigned HBASE-812:
---------------------------

    Assignee: Daniel Leffel

> Compaction needs little better skip algo
> ----------------------------------------
>
>                 Key: HBASE-812
>                 URL: https://issues.apache.org/jira/browse/HBASE-812
>             Project: Hadoop HBase
>          Issue Type: Bug
>          Components: regionserver
>    Affects Versions: 0.2.0, 0.3.0
>            Reporter: Billy Pearson
>            Assignee: Daniel Leffel
>            Priority: Critical
>             Fix For: 0.2.1, 0.3.0
>
>         Attachments: 812.patch.txt, 812v2.patch
>
>
> Looking at this section of one of my compaction's we have 3 files to compact the new algo is working great in my test but I see this below often we are skipping 2 out of the 3 files and compacting 1 file. 1 file is kind of a wast might as well just copy the file my suggestion is if there is only 1 file left after the new algo skips then just go on to the next column and skip the last file also. This will help improve compaction times a little more. 
> {code}
> 2008-08-10 10:00:45,310 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 1339600874/size: 4.6m, skipped 2, 4851776
> 2008-08-10 10:00:45,438 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 1 files into /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891
> 2008-08-10 10:00:46,838 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891 to /hbase/webdata/1339600874/size/mapfiles/7539342470259528578
> 2008-08-10 10:00:47,166 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 1339600874/size store size is 4.6m
> {code}

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Updated: (HBASE-812) Compaction needs little better skip algo

Posted by "stack (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HBASE-812?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

stack updated HBASE-812:
------------------------

    Priority: Critical  (was: Trivial)

Made this critical just because its so dumb that we're running compation of a single file.

> Compaction needs little better skip algo
> ----------------------------------------
>
>                 Key: HBASE-812
>                 URL: https://issues.apache.org/jira/browse/HBASE-812
>             Project: Hadoop HBase
>          Issue Type: Bug
>          Components: regionserver
>    Affects Versions: 0.2.0, 0.3.0
>            Reporter: Billy Pearson
>            Priority: Critical
>             Fix For: 0.2.1, 0.3.0
>
>
> Looking at this section of one of my compaction's we have 3 files to compact the new algo is working great in my test but I see this below often we are skipping 2 out of the 3 files and compacting 1 file. 1 file is kind of a wast might as well just copy the file my suggestion is if there is only 1 file left after the new algo skips then just go on to the next column and skip the last file also. This will help improve compaction times a little more. 
> {code}
> 2008-08-10 10:00:45,310 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 1339600874/size: 4.6m, skipped 2, 4851776
> 2008-08-10 10:00:45,438 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 1 files into /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891
> 2008-08-10 10:00:46,838 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891 to /hbase/webdata/1339600874/size/mapfiles/7539342470259528578
> 2008-08-10 10:00:47,166 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 1339600874/size store size is 4.6m
> {code}

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Commented: (HBASE-812) Compaction needs little better skip algo

Posted by "stack (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HBASE-812?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12621575#action_12621575 ] 

stack commented on HBASE-812:
-----------------------------

Billy: If we're compacting a single file, thats just dumb and we need to fix it (Doesn't look like the file we're compacting is a reference).  Do you have other examples?  Can you update to later hbase too -- looks like you are on older stuff -- and if the problem still exists, can you put in more examples?

> Compaction needs little better skip algo
> ----------------------------------------
>
>                 Key: HBASE-812
>                 URL: https://issues.apache.org/jira/browse/HBASE-812
>             Project: Hadoop HBase
>          Issue Type: Bug
>          Components: regionserver
>    Affects Versions: 0.2.0, 0.3.0
>            Reporter: Billy Pearson
>            Priority: Trivial
>             Fix For: 0.2.1, 0.3.0
>
>
> Looking at this section of one of my compaction's we have 3 files to compact the new algo is working great in my test but I see this below often we are skipping 2 out of the 3 files and compacting 1 file. 1 file is kind of a wast might as well just copy the file my suggestion is if there is only 1 file left after the new algo skips then just go on to the next column and skip the last file also. This will help improve compaction times a little more. 
> {code}
> 2008-08-10 10:00:45,310 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 1339600874/size: 4.6m, skipped 2, 4851776
> 2008-08-10 10:00:45,438 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 1 files into /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891
> 2008-08-10 10:00:46,838 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891 to /hbase/webdata/1339600874/size/mapfiles/7539342470259528578
> 2008-08-10 10:00:47,166 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 1339600874/size store size is 4.6m
> {code}

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Updated: (HBASE-812) Compaction needs little better skip algo

Posted by "stack (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HBASE-812?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

stack updated HBASE-812:
------------------------

    Resolution: Fixed
        Status: Resolved  (was: Patch Available)

Committed.  Thanks for the patch Daniel (and the edit Billy).

> Compaction needs little better skip algo
> ----------------------------------------
>
>                 Key: HBASE-812
>                 URL: https://issues.apache.org/jira/browse/HBASE-812
>             Project: Hadoop HBase
>          Issue Type: Bug
>          Components: regionserver
>    Affects Versions: 0.2.0, 0.3.0
>            Reporter: Billy Pearson
>            Assignee: Daniel Leffel
>            Priority: Critical
>             Fix For: 0.2.1, 0.3.0
>
>         Attachments: 812.patch.txt, 812v2.patch
>
>
> Looking at this section of one of my compaction's we have 3 files to compact the new algo is working great in my test but I see this below often we are skipping 2 out of the 3 files and compacting 1 file. 1 file is kind of a wast might as well just copy the file my suggestion is if there is only 1 file left after the new algo skips then just go on to the next column and skip the last file also. This will help improve compaction times a little more. 
> {code}
> 2008-08-10 10:00:45,310 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 1339600874/size: 4.6m, skipped 2, 4851776
> 2008-08-10 10:00:45,438 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 1 files into /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891
> 2008-08-10 10:00:46,838 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891 to /hbase/webdata/1339600874/size/mapfiles/7539342470259528578
> 2008-08-10 10:00:47,166 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 1339600874/size store size is 4.6m
> {code}

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Updated: (HBASE-812) Compaction needs little better skip algo

Posted by "Daniel Leffel (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HBASE-812?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Daniel Leffel updated HBASE-812:
--------------------------------

    Attachment: 812.patch.txt

Looks like this might fix the issue. It's my first patch though, so Billy, you might test it first.

> Compaction needs little better skip algo
> ----------------------------------------
>
>                 Key: HBASE-812
>                 URL: https://issues.apache.org/jira/browse/HBASE-812
>             Project: Hadoop HBase
>          Issue Type: Bug
>          Components: regionserver
>    Affects Versions: 0.2.0, 0.3.0
>            Reporter: Billy Pearson
>            Priority: Critical
>             Fix For: 0.2.1, 0.3.0
>
>         Attachments: 812.patch.txt
>
>
> Looking at this section of one of my compaction's we have 3 files to compact the new algo is working great in my test but I see this below often we are skipping 2 out of the 3 files and compacting 1 file. 1 file is kind of a wast might as well just copy the file my suggestion is if there is only 1 file left after the new algo skips then just go on to the next column and skip the last file also. This will help improve compaction times a little more. 
> {code}
> 2008-08-10 10:00:45,310 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 1339600874/size: 4.6m, skipped 2, 4851776
> 2008-08-10 10:00:45,438 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 1 files into /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891
> 2008-08-10 10:00:46,838 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891 to /hbase/webdata/1339600874/size/mapfiles/7539342470259528578
> 2008-08-10 10:00:47,166 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 1339600874/size store size is 4.6m
> {code}

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


[jira] Updated: (HBASE-812) Compaction needs little better skip algo

Posted by "Billy Pearson (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HBASE-812?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Billy Pearson updated HBASE-812:
--------------------------------

    Attachment: 812v2.patch

> Compaction needs little better skip algo
> ----------------------------------------
>
>                 Key: HBASE-812
>                 URL: https://issues.apache.org/jira/browse/HBASE-812
>             Project: Hadoop HBase
>          Issue Type: Bug
>          Components: regionserver
>    Affects Versions: 0.2.0, 0.3.0
>            Reporter: Billy Pearson
>            Priority: Critical
>             Fix For: 0.2.1, 0.3.0
>
>         Attachments: 812.patch.txt, 812v2.patch
>
>
> Looking at this section of one of my compaction's we have 3 files to compact the new algo is working great in my test but I see this below often we are skipping 2 out of the 3 files and compacting 1 file. 1 file is kind of a wast might as well just copy the file my suggestion is if there is only 1 file left after the new algo skips then just go on to the next column and skip the last file also. This will help improve compaction times a little more. 
> {code}
> 2008-08-10 10:00:45,310 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Compaction size of 1339600874/size: 4.6m, skipped 2, 4851776
> 2008-08-10 10:00:45,438 DEBUG org.apache.hadoop.hbase.regionserver.HStore: started compaction of 1 files into /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891
> 2008-08-10 10:00:46,838 DEBUG org.apache.hadoop.hbase.regionserver.HStore: moving /hbase/webdata/compaction.dir/1339600874/size/mapfiles/8653208152776334891 to /hbase/webdata/1339600874/size/mapfiles/7539342470259528578
> 2008-08-10 10:00:47,166 DEBUG org.apache.hadoop.hbase.regionserver.HStore: Completed compaction of 1339600874/size store size is 4.6m
> {code}

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.