You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "Roman Shaposhnik (Created) (JIRA)" <ji...@apache.org> on 2011/10/01 01:07:46 UTC

[jira] [Created] (HBASE-4526) special case for stopping master in hbase-daemon.sh is no longer required

special case for stopping master in hbase-daemon.sh is no longer required
-------------------------------------------------------------------------

                 Key: HBASE-4526
                 URL: https://issues.apache.org/jira/browse/HBASE-4526
             Project: HBase
          Issue Type: Bug
          Components: scripts
    Affects Versions: 0.90.4
            Reporter: Roman Shaposhnik
            Assignee: Roman Shaposhnik
            Priority: Minor


Now that HBASE-4209 is finally done (many thanks to stack for help and encouragement!) I don't think there's any reason to keep this bit of code in hbase-daemon.sh:

{noformat}
        if [ "$command" = "master" ]; then
          echo "`date` Killing $command" >> $loglog
          kill -9 `cat $pid` > /dev/null 2>&1
        else
          echo "`date` Killing $command" >> $loglog
          kill `cat $pid` > /dev/null 2>&1
        fi
{noformat}

I suggest we remove the special case completely, since I don't think it serves any useful purpose (patch attached).

Now, as an additional precautionary measure. We can try to follow up with a SIGKILL once a certain timeout expires.
Let me know if you think it is necessary to do so and I'll update the patch.

--
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] (HBASE-4526) special case for stopping master in hbase-daemon.sh is no longer required

Posted by "stack (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HBASE-4526?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13118670#comment-13118670 ] 

stack commented on HBASE-4526:
------------------------------

Oh, I could try it for you if you don't have a little old distributed cluster to go against.  let me know.
                
> special case for stopping master in hbase-daemon.sh is no longer required
> -------------------------------------------------------------------------
>
>                 Key: HBASE-4526
>                 URL: https://issues.apache.org/jira/browse/HBASE-4526
>             Project: HBase
>          Issue Type: Bug
>          Components: scripts
>    Affects Versions: 0.90.4
>            Reporter: Roman Shaposhnik
>            Assignee: Roman Shaposhnik
>            Priority: Minor
>         Attachments: HBASE-4526.patch.txt
>
>
> Now that HBASE-4209 is finally done (many thanks to stack for help and encouragement!) I don't think there's any reason to keep this bit of code in hbase-daemon.sh:
> {noformat}
>         if [ "$command" = "master" ]; then
>           echo "`date` Killing $command" >> $loglog
>           kill -9 `cat $pid` > /dev/null 2>&1
>         else
>           echo "`date` Killing $command" >> $loglog
>           kill `cat $pid` > /dev/null 2>&1
>         fi
> {noformat}
> I suggest we remove the special case completely, since I don't think it serves any useful purpose (patch attached).
> Now, as an additional precautionary measure. We can try to follow up with a SIGKILL once a certain timeout expires.
> Let me know if you think it is necessary to do so and I'll update the patch.

--
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] (HBASE-4526) special case for stopping master in hbase-daemon.sh is no longer required

Posted by "Roman Shaposhnik (Updated) (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HBASE-4526?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Roman Shaposhnik updated HBASE-4526:
------------------------------------

    Attachment: HBASE-4526.patch.txt
    
> special case for stopping master in hbase-daemon.sh is no longer required
> -------------------------------------------------------------------------
>
>                 Key: HBASE-4526
>                 URL: https://issues.apache.org/jira/browse/HBASE-4526
>             Project: HBase
>          Issue Type: Bug
>          Components: scripts
>    Affects Versions: 0.90.4
>            Reporter: Roman Shaposhnik
>            Assignee: Roman Shaposhnik
>            Priority: Minor
>         Attachments: HBASE-4526.patch.txt
>
>
> Now that HBASE-4209 is finally done (many thanks to stack for help and encouragement!) I don't think there's any reason to keep this bit of code in hbase-daemon.sh:
> {noformat}
>         if [ "$command" = "master" ]; then
>           echo "`date` Killing $command" >> $loglog
>           kill -9 `cat $pid` > /dev/null 2>&1
>         else
>           echo "`date` Killing $command" >> $loglog
>           kill `cat $pid` > /dev/null 2>&1
>         fi
> {noformat}
> I suggest we remove the special case completely, since I don't think it serves any useful purpose (patch attached).
> Now, as an additional precautionary measure. We can try to follow up with a SIGKILL once a certain timeout expires.
> Let me know if you think it is necessary to do so and I'll update the patch.

--
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] (HBASE-4526) special case for stopping master in hbase-daemon.sh is no longer required

Posted by "Roman Shaposhnik (Updated) (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HBASE-4526?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Roman Shaposhnik updated HBASE-4526:
------------------------------------

    Status: Patch Available  (was: Open)
    
> special case for stopping master in hbase-daemon.sh is no longer required
> -------------------------------------------------------------------------
>
>                 Key: HBASE-4526
>                 URL: https://issues.apache.org/jira/browse/HBASE-4526
>             Project: HBase
>          Issue Type: Bug
>          Components: scripts
>    Affects Versions: 0.90.4
>            Reporter: Roman Shaposhnik
>            Assignee: Roman Shaposhnik
>            Priority: Minor
>         Attachments: HBASE-4526.patch.txt
>
>
> Now that HBASE-4209 is finally done (many thanks to stack for help and encouragement!) I don't think there's any reason to keep this bit of code in hbase-daemon.sh:
> {noformat}
>         if [ "$command" = "master" ]; then
>           echo "`date` Killing $command" >> $loglog
>           kill -9 `cat $pid` > /dev/null 2>&1
>         else
>           echo "`date` Killing $command" >> $loglog
>           kill `cat $pid` > /dev/null 2>&1
>         fi
> {noformat}
> I suggest we remove the special case completely, since I don't think it serves any useful purpose (patch attached).
> Now, as an additional precautionary measure. We can try to follow up with a SIGKILL once a certain timeout expires.
> Let me know if you think it is necessary to do so and I'll update the patch.

--
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] (HBASE-4526) special case for stopping master in hbase-daemon.sh is no longer required

Posted by "Roman Shaposhnik (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HBASE-4526?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13118938#comment-13118938 ] 

Roman Shaposhnik commented on HBASE-4526:
-----------------------------------------

Actually, the only thing that my patch changes is what signal gets used to stop master. Currently it is SIGKILL which is very harsh. And after my change it'll be SIGTERM which gives JVM a chance to shutdown. No other changes in behavior.

More to the point -- I tried this in a fully distributed case (master and regionserver running on different nodes) and it seems that I can stop the master without affecting the regionserver.
                
> special case for stopping master in hbase-daemon.sh is no longer required
> -------------------------------------------------------------------------
>
>                 Key: HBASE-4526
>                 URL: https://issues.apache.org/jira/browse/HBASE-4526
>             Project: HBase
>          Issue Type: Bug
>          Components: scripts
>    Affects Versions: 0.90.4
>            Reporter: Roman Shaposhnik
>            Assignee: Roman Shaposhnik
>            Priority: Minor
>         Attachments: HBASE-4526.patch.txt
>
>
> Now that HBASE-4209 is finally done (many thanks to stack for help and encouragement!) I don't think there's any reason to keep this bit of code in hbase-daemon.sh:
> {noformat}
>         if [ "$command" = "master" ]; then
>           echo "`date` Killing $command" >> $loglog
>           kill -9 `cat $pid` > /dev/null 2>&1
>         else
>           echo "`date` Killing $command" >> $loglog
>           kill `cat $pid` > /dev/null 2>&1
>         fi
> {noformat}
> I suggest we remove the special case completely, since I don't think it serves any useful purpose (patch attached).
> Now, as an additional precautionary measure. We can try to follow up with a SIGKILL once a certain timeout expires.
> Let me know if you think it is necessary to do so and I'll update the patch.

--
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] (HBASE-4526) special case for stopping master in hbase-daemon.sh is no longer required

Posted by "stack (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HBASE-4526?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13118669#comment-13118669 ] 

stack commented on HBASE-4526:
------------------------------

Have you tried this on a cluster Roman and your change does not cause the whole cluster to go down?  It just shuts down the master?

(Your patch makes sense but I have the above nagging feeling we're changing what the command does).
                
> special case for stopping master in hbase-daemon.sh is no longer required
> -------------------------------------------------------------------------
>
>                 Key: HBASE-4526
>                 URL: https://issues.apache.org/jira/browse/HBASE-4526
>             Project: HBase
>          Issue Type: Bug
>          Components: scripts
>    Affects Versions: 0.90.4
>            Reporter: Roman Shaposhnik
>            Assignee: Roman Shaposhnik
>            Priority: Minor
>         Attachments: HBASE-4526.patch.txt
>
>
> Now that HBASE-4209 is finally done (many thanks to stack for help and encouragement!) I don't think there's any reason to keep this bit of code in hbase-daemon.sh:
> {noformat}
>         if [ "$command" = "master" ]; then
>           echo "`date` Killing $command" >> $loglog
>           kill -9 `cat $pid` > /dev/null 2>&1
>         else
>           echo "`date` Killing $command" >> $loglog
>           kill `cat $pid` > /dev/null 2>&1
>         fi
> {noformat}
> I suggest we remove the special case completely, since I don't think it serves any useful purpose (patch attached).
> Now, as an additional precautionary measure. We can try to follow up with a SIGKILL once a certain timeout expires.
> Let me know if you think it is necessary to do so and I'll update the patch.

--
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] (HBASE-4526) special case for stopping master in hbase-daemon.sh is no longer required

Posted by "stack (Updated) (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HBASE-4526?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

stack updated HBASE-4526:
-------------------------

       Resolution: Fixed
    Fix Version/s: 0.92.0
     Hadoop Flags: Reviewed
           Status: Resolved  (was: Patch Available)

Committed to 0.92 branch and trunk.  Thank you for the patch Roman.
                
> special case for stopping master in hbase-daemon.sh is no longer required
> -------------------------------------------------------------------------
>
>                 Key: HBASE-4526
>                 URL: https://issues.apache.org/jira/browse/HBASE-4526
>             Project: HBase
>          Issue Type: Bug
>          Components: scripts
>    Affects Versions: 0.90.4
>            Reporter: Roman Shaposhnik
>            Assignee: Roman Shaposhnik
>            Priority: Minor
>             Fix For: 0.92.0
>
>         Attachments: HBASE-4526.patch.txt
>
>
> Now that HBASE-4209 is finally done (many thanks to stack for help and encouragement!) I don't think there's any reason to keep this bit of code in hbase-daemon.sh:
> {noformat}
>         if [ "$command" = "master" ]; then
>           echo "`date` Killing $command" >> $loglog
>           kill -9 `cat $pid` > /dev/null 2>&1
>         else
>           echo "`date` Killing $command" >> $loglog
>           kill `cat $pid` > /dev/null 2>&1
>         fi
> {noformat}
> I suggest we remove the special case completely, since I don't think it serves any useful purpose (patch attached).
> Now, as an additional precautionary measure. We can try to follow up with a SIGKILL once a certain timeout expires.
> Let me know if you think it is necessary to do so and I'll update the patch.

--
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] (HBASE-4526) special case for stopping master in hbase-daemon.sh is no longer required

Posted by "Hudson (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HBASE-4526?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13119515#comment-13119515 ] 

Hudson commented on HBASE-4526:
-------------------------------

Integrated in HBase-0.92 #39 (See [https://builds.apache.org/job/HBase-0.92/39/])
    HBASE-4526 special case for stopping master in hbase-daemon.sh is no longer required

stack : 
Files : 
* /hbase/branches/0.92/CHANGES.txt
* /hbase/branches/0.92/bin/hbase-daemon.sh

                
> special case for stopping master in hbase-daemon.sh is no longer required
> -------------------------------------------------------------------------
>
>                 Key: HBASE-4526
>                 URL: https://issues.apache.org/jira/browse/HBASE-4526
>             Project: HBase
>          Issue Type: Bug
>          Components: scripts
>    Affects Versions: 0.90.4
>            Reporter: Roman Shaposhnik
>            Assignee: Roman Shaposhnik
>            Priority: Minor
>             Fix For: 0.92.0
>
>         Attachments: HBASE-4526.patch.txt
>
>
> Now that HBASE-4209 is finally done (many thanks to stack for help and encouragement!) I don't think there's any reason to keep this bit of code in hbase-daemon.sh:
> {noformat}
>         if [ "$command" = "master" ]; then
>           echo "`date` Killing $command" >> $loglog
>           kill -9 `cat $pid` > /dev/null 2>&1
>         else
>           echo "`date` Killing $command" >> $loglog
>           kill `cat $pid` > /dev/null 2>&1
>         fi
> {noformat}
> I suggest we remove the special case completely, since I don't think it serves any useful purpose (patch attached).
> Now, as an additional precautionary measure. We can try to follow up with a SIGKILL once a certain timeout expires.
> Let me know if you think it is necessary to do so and I'll update the patch.

--
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