You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@accumulo.apache.org by "Keith Turner (Created) (JIRA)" <ji...@apache.org> on 2012/02/14 18:03:59 UTC
[jira] [Created] (ACCUMULO-398) Table tablets not evenly spread.
Table tablets not evenly spread.
--------------------------------
Key: ACCUMULO-398
URL: https://issues.apache.org/jira/browse/ACCUMULO-398
Project: Accumulo
Issue Type: Bug
Reporter: Keith Turner
Assignee: Eric Newton
Fix For: 1.4.0
I was running random walk test (with agitation) and wanted to switch to running continuous ingest test. So I killed the random walk test (left all of its tables), created a continuous ingest table w/ 16 tablets, and started continuous ingest. I noticed that the 16 tablets were not evenly spread across the 10 tablet servers. 2 tservers had 0, 2 tservers had 1, 5 tservers had 2, 1 tserver had 4. I do not think this is caused by ACCUMULO-393 because I keep seeing the following in the logs.
{noformat}
14 17:02:03,325 [balancer.DefaultLoadBalancer] DEBUG: balance ended with 0 migrations
{noformat}
--
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-398) Table tablets not evenly spread.
Posted by "Keith Turner (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-398?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13208688#comment-13208688 ]
Keith Turner commented on ACCUMULO-398:
---------------------------------------
The current code is probably working as expected. I was under the impression that the default Accumulo configuration would evenly balance each tables tablets across tablet servers. However this is not the default behavior. The default behavior is to spread all tablets in the system evenly across tablet servers. To spread table tablets evenly the configuration master.tablet.balancer would need to be set to org.apache.accumulo.server.master.balancer.TableLoadBalancer. This should probably be the system default. However I am not sure I am comfortable making this change for 1.4.0 w/ all of the testing that has already been done. Maybe this change could be made for 1.4.1?
Oh when I saw the CI tables tablets were not evenly balanced, I did not notice that all tablets were evenly balanced. Each tablet server had about the same # of tablets.
> Table tablets not evenly spread.
> --------------------------------
>
> Key: ACCUMULO-398
> URL: https://issues.apache.org/jira/browse/ACCUMULO-398
> Project: Accumulo
> Issue Type: Bug
> Reporter: Keith Turner
> Assignee: Eric Newton
> Labels: 14_qa_bug
> Fix For: 1.4.0
>
>
> I was running random walk test (with agitation) and wanted to switch to running continuous ingest test. So I killed the random walk test (left all of its tables), created a continuous ingest table w/ 16 tablets, and started continuous ingest. I noticed that the 16 tablets were not evenly spread across the 10 tablet servers. 2 tservers had 0, 2 tservers had 1, 5 tservers had 2, 1 tserver had 4. I do not think this is caused by ACCUMULO-393 because I keep seeing the following in the logs.
> {noformat}
> 14 17:02:03,325 [balancer.DefaultLoadBalancer] DEBUG: balance ended with 0 migrations
> {noformat}
--
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-398) Table tablets not evenly spread.
Posted by "Keith Turner (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-398?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13207835#comment-13207835 ]
Keith Turner commented on ACCUMULO-398:
---------------------------------------
Tried restartig the master and this did not help. I deleted all of the other tables and then it spread the tablets out evenly.
> Table tablets not evenly spread.
> --------------------------------
>
> Key: ACCUMULO-398
> URL: https://issues.apache.org/jira/browse/ACCUMULO-398
> Project: Accumulo
> Issue Type: Bug
> Reporter: Keith Turner
> Assignee: Eric Newton
> Labels: 14_qa_bug
> Fix For: 1.4.0
>
>
> I was running random walk test (with agitation) and wanted to switch to running continuous ingest test. So I killed the random walk test (left all of its tables), created a continuous ingest table w/ 16 tablets, and started continuous ingest. I noticed that the 16 tablets were not evenly spread across the 10 tablet servers. 2 tservers had 0, 2 tservers had 1, 5 tservers had 2, 1 tserver had 4. I do not think this is caused by ACCUMULO-393 because I keep seeing the following in the logs.
> {noformat}
> 14 17:02:03,325 [balancer.DefaultLoadBalancer] DEBUG: balance ended with 0 migrations
> {noformat}
--
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] [Updated] (ACCUMULO-398) Table tablets not evenly spread.
Posted by "Keith Turner (Updated) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/ACCUMULO-398?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Keith Turner updated ACCUMULO-398:
----------------------------------
Fix Version/s: (was: 1.4.0)
1.4.1
> Table tablets not evenly spread.
> --------------------------------
>
> Key: ACCUMULO-398
> URL: https://issues.apache.org/jira/browse/ACCUMULO-398
> Project: Accumulo
> Issue Type: Bug
> Reporter: Keith Turner
> Assignee: Eric Newton
> Labels: 14_qa_bug
> Fix For: 1.4.1
>
>
> I was running random walk test (with agitation) and wanted to switch to running continuous ingest test. So I killed the random walk test (left all of its tables), created a continuous ingest table w/ 16 tablets, and started continuous ingest. I noticed that the 16 tablets were not evenly spread across the 10 tablet servers. 2 tservers had 0, 2 tservers had 1, 5 tservers had 2, 1 tserver had 4. I do not think this is caused by ACCUMULO-393 because I keep seeing the following in the logs.
> {noformat}
> 14 17:02:03,325 [balancer.DefaultLoadBalancer] DEBUG: balance ended with 0 migrations
> {noformat}
--
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