You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@accumulo.apache.org by "John Vines (Created) (JIRA)" <ji...@apache.org> on 2011/11/15 17:58:52 UTC
[jira] [Created] (ACCUMULO-145) Release Accumulo 1.4
Release Accumulo 1.4
--------------------
Key: ACCUMULO-145
URL: https://issues.apache.org/jira/browse/ACCUMULO-145
Project: Accumulo
Issue Type: Task
Reporter: John Vines
Priority: Blocker
Fix For: 1.4.0
The 1.4 branch of Accumulo is preparing release. This ticket should take us through the release process. On the list of things left to do are:
1. Test as per http://incubator.apache.org/accumulo/governance/releasing.html
2. Document the release, including the documentation that needs to done according to Accumulo-68
3. Purge GPL content as per Accumulo-92
3. Tag the code
4. Do the build
5. Sign the build
6. Take a vote
See also http://incubator.apache.org/guides/releasemanagement.html , http://incubator.apache.org/accumulo/governance/releasing.html
--
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
[jira] [Commented] (ACCUMULO-145) Release Accumulo 1.4
Posted by "Keith Turner (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13207171#comment-13207171 ]
Keith Turner commented on ACCUMULO-145:
---------------------------------------
I was able to determine why the verification map reduce job was running slowly. Also, in an earlier post I thought there was a network issue causing a slow node. This was not the case, a tserver was backed up on scans but it was not caused by a network issues. I misread .223ms as 223ms.
Issues ACCUMULO-393 and ACCUMULO-395 were partially to blame, but that was not the whole story. To get around ACCUMULO-393 I restarted the master. After the restart, all tablets were balanced. Then I ran the verification map reduce job again. I noticed that scans still backed up on some tservers. Looking into this I noticed ACCUMULO-395. So map task were not running locally and more importantly they were not accessing tablet servers uniformly. There were 1024 tablets in the continuous ingest table. Tablets had an average of 8.2 files with a standard deviation of 2.3. So lots of clients reading from lots of tablets on a tablet server with each tablet having lots of files caused thrashing. Once a tablet server started thrashing, the mappers running against other tservers would finish quickly leading to more mappers eventually going to the thrashing tserver. I made tablets compact down to four files (by setting table.file.max) and the map reduce job ran in 2h12m as opposed to 21h13m. So in short ACCUMULO-393, ACCUMULO-395, and too many files per tablet made the job run super slow.
> Release Accumulo 1.4
> --------------------
>
> Key: ACCUMULO-145
> URL: https://issues.apache.org/jira/browse/ACCUMULO-145
> Project: Accumulo
> Issue Type: Task
> Reporter: John Vines
> Assignee: Eric Newton
> Priority: Blocker
> Labels: release
> Fix For: 1.4.0
>
>
> The 1.4 branch of Accumulo is preparing release. This ticket should take us through the release process. On the list of things left to do are:
> 1. Test as per http://incubator.apache.org/accumulo/governance/releasing.html
> 2. Document the release, including the documentation that needs to done according to Accumulo-68
> 3. Purge GPL content as per Accumulo-92
> 3. Tag the code
> 4. Do the build
> 5. Sign the build
> 6. Take a vote
> See also http://incubator.apache.org/guides/releasemanagement.html , http://incubator.apache.org/accumulo/governance/releasing.html
--
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
[jira] [Commented] (ACCUMULO-145) Release Accumulo 1.4
Posted by "Keith Turner (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13206937#comment-13206937 ]
Keith Turner commented on ACCUMULO-145:
---------------------------------------
Completed a round of continuous ingest on a 9 node cluster w/ agitation. Verification succeeded, no data was lost. The map reduce verification job ran very slowly, at least some task did. Some task completed in a few minutes, some took a few hours. All had the same amount of data. Not sure what happened.
Ingest ran for 14 hours. The agitator killed loggers 285 times, tablet servers 264 times, and the master 55 times. 16 billion entries were ingested.
> Release Accumulo 1.4
> --------------------
>
> Key: ACCUMULO-145
> URL: https://issues.apache.org/jira/browse/ACCUMULO-145
> Project: Accumulo
> Issue Type: Task
> Reporter: John Vines
> Assignee: Eric Newton
> Priority: Blocker
> Labels: release
> Fix For: 1.4.0
>
>
> The 1.4 branch of Accumulo is preparing release. This ticket should take us through the release process. On the list of things left to do are:
> 1. Test as per http://incubator.apache.org/accumulo/governance/releasing.html
> 2. Document the release, including the documentation that needs to done according to Accumulo-68
> 3. Purge GPL content as per Accumulo-92
> 3. Tag the code
> 4. Do the build
> 5. Sign the build
> 6. Take a vote
> See also http://incubator.apache.org/guides/releasemanagement.html , http://incubator.apache.org/accumulo/governance/releasing.html
--
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
[jira] [Commented] (ACCUMULO-145) Release Accumulo 1.4
Posted by "Keith Turner (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13197957#comment-13197957 ]
Keith Turner commented on ACCUMULO-145:
---------------------------------------
MAde it through a run of the random walk test last night w/o agitation and w/o the security random walk test. The security random walk test keeps dying on some known issues that will not be fixed in 1.4. Ran on a 10 node cluster w/ 10 walkers. Three of the walkers died w/ a know issue ACCUMULO-357. I think this is the first successful (no new issues found) overnight run.
> Release Accumulo 1.4
> --------------------
>
> Key: ACCUMULO-145
> URL: https://issues.apache.org/jira/browse/ACCUMULO-145
> Project: Accumulo
> Issue Type: Task
> Reporter: John Vines
> Assignee: Eric Newton
> Priority: Blocker
> Labels: release
> Fix For: 1.4.0
>
>
> The 1.4 branch of Accumulo is preparing release. This ticket should take us through the release process. On the list of things left to do are:
> 1. Test as per http://incubator.apache.org/accumulo/governance/releasing.html
> 2. Document the release, including the documentation that needs to done according to Accumulo-68
> 3. Purge GPL content as per Accumulo-92
> 3. Tag the code
> 4. Do the build
> 5. Sign the build
> 6. Take a vote
> See also http://incubator.apache.org/guides/releasemanagement.html , http://incubator.apache.org/accumulo/governance/releasing.html
--
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
[jira] [Commented] (ACCUMULO-145) Release Accumulo 1.4
Posted by "Keith Turner (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13220092#comment-13220092 ]
Keith Turner commented on ACCUMULO-145:
---------------------------------------
Random walk (w/ agitation) ran overnight w/o issue.
> Release Accumulo 1.4
> --------------------
>
> Key: ACCUMULO-145
> URL: https://issues.apache.org/jira/browse/ACCUMULO-145
> Project: Accumulo
> Issue Type: Task
> Reporter: John Vines
> Assignee: Eric Newton
> Priority: Blocker
> Labels: release
> Fix For: 1.4.0
>
>
> The 1.4 branch of Accumulo is preparing release. This ticket should take us through the release process. On the list of things left to do are:
> 1. Test as per http://incubator.apache.org/accumulo/governance/releasing.html
> 2. Document the release, including the documentation that needs to done according to Accumulo-68
> 3. Purge GPL content as per Accumulo-92
> 3. Tag the code
> 4. Do the build
> 5. Sign the build
> 6. Take a vote
> See also http://incubator.apache.org/guides/releasemanagement.html , http://incubator.apache.org/accumulo/governance/releasing.html
--
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
[jira] [Commented] (ACCUMULO-145) Release Accumulo 1.4
Posted by "Keith Turner (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13205081#comment-13205081 ]
Keith Turner commented on ACCUMULO-145:
---------------------------------------
I was curious how many bugs we found in 1.4 testing so far. So I looked through the tickets and started labeling bugs found during 1.4 testing with 14_qa_bug. I am not sure if I labeled everything.
[Bugs found in 1.4 testing|https://issues.apache.org/jira/secure/IssueNavigator.jspa?reset=true&jqlQuery=labels+%3D+14_qa_bug]
> Release Accumulo 1.4
> --------------------
>
> Key: ACCUMULO-145
> URL: https://issues.apache.org/jira/browse/ACCUMULO-145
> Project: Accumulo
> Issue Type: Task
> Reporter: John Vines
> Assignee: Eric Newton
> Priority: Blocker
> Labels: release
> Fix For: 1.4.0
>
>
> The 1.4 branch of Accumulo is preparing release. This ticket should take us through the release process. On the list of things left to do are:
> 1. Test as per http://incubator.apache.org/accumulo/governance/releasing.html
> 2. Document the release, including the documentation that needs to done according to Accumulo-68
> 3. Purge GPL content as per Accumulo-92
> 3. Tag the code
> 4. Do the build
> 5. Sign the build
> 6. Take a vote
> See also http://incubator.apache.org/guides/releasemanagement.html , http://incubator.apache.org/accumulo/governance/releasing.html
--
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
[jira] [Commented] (ACCUMULO-145) Release Accumulo 1.4
Posted by "Keith Turner (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13212701#comment-13212701 ]
Keith Turner commented on ACCUMULO-145:
---------------------------------------
Random walk (w/o agitation) ran over a three day weekend on our 10 node test cluster w/ no issues except for one tablet server lost because of ACCUMULO-419.
> Release Accumulo 1.4
> --------------------
>
> Key: ACCUMULO-145
> URL: https://issues.apache.org/jira/browse/ACCUMULO-145
> Project: Accumulo
> Issue Type: Task
> Reporter: John Vines
> Assignee: Eric Newton
> Priority: Blocker
> Labels: release
> Fix For: 1.4.0
>
>
> The 1.4 branch of Accumulo is preparing release. This ticket should take us through the release process. On the list of things left to do are:
> 1. Test as per http://incubator.apache.org/accumulo/governance/releasing.html
> 2. Document the release, including the documentation that needs to done according to Accumulo-68
> 3. Purge GPL content as per Accumulo-92
> 3. Tag the code
> 4. Do the build
> 5. Sign the build
> 6. Take a vote
> See also http://incubator.apache.org/guides/releasemanagement.html , http://incubator.apache.org/accumulo/governance/releasing.html
--
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
[jira] [Commented] (ACCUMULO-145) Release Accumulo 1.4
Posted by "Keith Turner (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13204930#comment-13204930 ]
Keith Turner commented on ACCUMULO-145:
---------------------------------------
MAde it through a 21 hour run of the random walk test last w/o agitation and w/o the security random walk test. No new issues occurred.
> Release Accumulo 1.4
> --------------------
>
> Key: ACCUMULO-145
> URL: https://issues.apache.org/jira/browse/ACCUMULO-145
> Project: Accumulo
> Issue Type: Task
> Reporter: John Vines
> Assignee: Eric Newton
> Priority: Blocker
> Labels: release
> Fix For: 1.4.0
>
>
> The 1.4 branch of Accumulo is preparing release. This ticket should take us through the release process. On the list of things left to do are:
> 1. Test as per http://incubator.apache.org/accumulo/governance/releasing.html
> 2. Document the release, including the documentation that needs to done according to Accumulo-68
> 3. Purge GPL content as per Accumulo-92
> 3. Tag the code
> 4. Do the build
> 5. Sign the build
> 6. Take a vote
> See also http://incubator.apache.org/guides/releasemanagement.html , http://incubator.apache.org/accumulo/governance/releasing.html
--
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
[jira] [Commented] (ACCUMULO-145) Release Accumulo 1.4
Posted by "Keith Turner (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13206169#comment-13206169 ]
Keith Turner commented on ACCUMULO-145:
---------------------------------------
Completed a round of continuous ingest on a 10 node cluster w/o agitation. Verification succeeded, no data was lost. With the new running scans feature on the monitor page I noticed one node was slow during verification. It had queued scans when no other node did. Finally all mappers were waiting on it. Not sure why it was slow yet, does not appear to be an Accumulo bug because pings are taking hundreds of milliseconds. This probably slowed ingest too, but the slowest node will slow all other nodes during ingest unlike verification so its not as easily detectable on the monitor page. During ingest clients write to all nodes, during verification clients mappers read individual tablets. For this run of about 14.5 hours 18.2 billion nodes were ingested and nothing was lost. Wonder if it would have gone faster w/o the slow node. Next I will run w/ agitation
> Release Accumulo 1.4
> --------------------
>
> Key: ACCUMULO-145
> URL: https://issues.apache.org/jira/browse/ACCUMULO-145
> Project: Accumulo
> Issue Type: Task
> Reporter: John Vines
> Assignee: Eric Newton
> Priority: Blocker
> Labels: release
> Fix For: 1.4.0
>
>
> The 1.4 branch of Accumulo is preparing release. This ticket should take us through the release process. On the list of things left to do are:
> 1. Test as per http://incubator.apache.org/accumulo/governance/releasing.html
> 2. Document the release, including the documentation that needs to done according to Accumulo-68
> 3. Purge GPL content as per Accumulo-92
> 3. Tag the code
> 4. Do the build
> 5. Sign the build
> 6. Take a vote
> See also http://incubator.apache.org/guides/releasemanagement.html , http://incubator.apache.org/accumulo/governance/releasing.html
--
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
[jira] [Commented] (ACCUMULO-145) Release Accumulo 1.4
Posted by "Keith Turner (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13207334#comment-13207334 ]
Keith Turner commented on ACCUMULO-145:
---------------------------------------
After putting in a fix for ACCUMULO-395 the verification job ran in 1h45m, a nice little improvement over 2h12m.
> Release Accumulo 1.4
> --------------------
>
> Key: ACCUMULO-145
> URL: https://issues.apache.org/jira/browse/ACCUMULO-145
> Project: Accumulo
> Issue Type: Task
> Reporter: John Vines
> Assignee: Eric Newton
> Priority: Blocker
> Labels: release
> Fix For: 1.4.0
>
>
> The 1.4 branch of Accumulo is preparing release. This ticket should take us through the release process. On the list of things left to do are:
> 1. Test as per http://incubator.apache.org/accumulo/governance/releasing.html
> 2. Document the release, including the documentation that needs to done according to Accumulo-68
> 3. Purge GPL content as per Accumulo-92
> 3. Tag the code
> 4. Do the build
> 5. Sign the build
> 6. Take a vote
> See also http://incubator.apache.org/guides/releasemanagement.html , http://incubator.apache.org/accumulo/governance/releasing.html
--
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
[jira] [Assigned] (ACCUMULO-145) Release Accumulo 1.4
Posted by "Eric Newton (Assigned) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-145?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Eric Newton reassigned ACCUMULO-145:
------------------------------------
Assignee: Eric Newton
> Release Accumulo 1.4
> --------------------
>
> Key: ACCUMULO-145
> URL: https://issues.apache.org/jira/browse/ACCUMULO-145
> Project: Accumulo
> Issue Type: Task
> Reporter: John Vines
> Assignee: Eric Newton
> Priority: Blocker
> Labels: release
> Fix For: 1.4.0
>
>
> The 1.4 branch of Accumulo is preparing release. This ticket should take us through the release process. On the list of things left to do are:
> 1. Test as per http://incubator.apache.org/accumulo/governance/releasing.html
> 2. Document the release, including the documentation that needs to done according to Accumulo-68
> 3. Purge GPL content as per Accumulo-92
> 3. Tag the code
> 4. Do the build
> 5. Sign the build
> 6. Take a vote
> See also http://incubator.apache.org/guides/releasemanagement.html , http://incubator.apache.org/accumulo/governance/releasing.html
--
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