You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "Jonathan Gray (JIRA)" <ji...@apache.org> on 2010/10/28 08:12:07 UTC

[jira] Created: (HBASE-3163) If we timeout PENDING_CLOSE and send another closeRegion RPC, need to handle NSRE from RS (comes as a RemoteException)

If we timeout PENDING_CLOSE and send another closeRegion RPC, need to handle NSRE from RS (comes as a RemoteException)
----------------------------------------------------------------------------------------------------------------------

                 Key: HBASE-3163
                 URL: https://issues.apache.org/jira/browse/HBASE-3163
             Project: HBase
          Issue Type: Bug
          Components: master
    Affects Versions: 0.90.0
            Reporter: Jonathan Gray
            Priority: Critical
             Fix For: 0.90.0


When we send a closeRegion RPC to an RS, we are catching NSRE but when the RS is the one throwing the NSRE, then it comes back as a RemoteException (then an NSRE) and we aren't unwrapping it properly.

We need to catch this and then deal with it appropriately.

Still tracking how this happened in the first place.

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

[jira] Updated: (HBASE-3163) If we timeout PENDING_CLOSE and send another closeRegion RPC, need to handle NSRE from RS (comes as a RemoteException)

Posted by "Jonathan Gray (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HBASE-3163?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jonathan Gray updated HBASE-3163:
---------------------------------

      Resolution: Fixed
    Hadoop Flags: [Reviewed]
          Status: Resolved  (was: Patch Available)

Thanks for review stack.  Committed to trunk.

> If we timeout PENDING_CLOSE and send another closeRegion RPC, need to handle NSRE from RS (comes as a RemoteException)
> ----------------------------------------------------------------------------------------------------------------------
>
>                 Key: HBASE-3163
>                 URL: https://issues.apache.org/jira/browse/HBASE-3163
>             Project: HBase
>          Issue Type: Bug
>          Components: master
>    Affects Versions: 0.90.0
>            Reporter: Jonathan Gray
>            Assignee: Jonathan Gray
>            Priority: Critical
>             Fix For: 0.90.0
>
>
> When we send a closeRegion RPC to an RS, we are catching NSRE but when the RS is the one throwing the NSRE, then it comes back as a RemoteException (then an NSRE) and we aren't unwrapping it properly.
> We need to catch this and then deal with it appropriately.
> Still tracking how this happened in the first place.

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

[jira] Commented: (HBASE-3163) If we timeout PENDING_CLOSE and send another closeRegion RPC, need to handle NSRE from RS (comes as a RemoteException)

Posted by "HBase Review Board (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HBASE-3163?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12926490#action_12926490 ] 

HBase Review Board commented on HBASE-3163:
-------------------------------------------

Message from: "Jonathan Gray" <jg...@apache.org>

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://review.cloudera.org/r/1130/
-----------------------------------------------------------

Review request for hbase and stack.


Summary
-------

Unwraps RemoteExceptions and we don't abort if it's an NSRE.

This handling is rather different from the handling of opens (we abort in some cases here), so we might consider making them the same and no longer aborting here.  I'm fine with leaving it for a while longer as we do all this testing.


This addresses bug HBASE-3163.
    http://issues.apache.org/jira/browse/HBASE-3163


Diffs
-----

  trunk/src/main/java/org/apache/hadoop/hbase/master/AssignmentManager.java 1028936 

Diff: http://review.cloudera.org/r/1130/diff


Testing
-------


Thanks,

Jonathan




> If we timeout PENDING_CLOSE and send another closeRegion RPC, need to handle NSRE from RS (comes as a RemoteException)
> ----------------------------------------------------------------------------------------------------------------------
>
>                 Key: HBASE-3163
>                 URL: https://issues.apache.org/jira/browse/HBASE-3163
>             Project: HBase
>          Issue Type: Bug
>          Components: master
>    Affects Versions: 0.90.0
>            Reporter: Jonathan Gray
>            Priority: Critical
>             Fix For: 0.90.0
>
>
> When we send a closeRegion RPC to an RS, we are catching NSRE but when the RS is the one throwing the NSRE, then it comes back as a RemoteException (then an NSRE) and we aren't unwrapping it properly.
> We need to catch this and then deal with it appropriately.
> Still tracking how this happened in the first place.

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

[jira] Updated: (HBASE-3163) If we timeout PENDING_CLOSE and send another closeRegion RPC, need to handle NSRE from RS (comes as a RemoteException)

Posted by "Jonathan Gray (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HBASE-3163?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jonathan Gray updated HBASE-3163:
---------------------------------

    Status: Patch Available  (was: Open)

> If we timeout PENDING_CLOSE and send another closeRegion RPC, need to handle NSRE from RS (comes as a RemoteException)
> ----------------------------------------------------------------------------------------------------------------------
>
>                 Key: HBASE-3163
>                 URL: https://issues.apache.org/jira/browse/HBASE-3163
>             Project: HBase
>          Issue Type: Bug
>          Components: master
>    Affects Versions: 0.90.0
>            Reporter: Jonathan Gray
>            Priority: Critical
>             Fix For: 0.90.0
>
>
> When we send a closeRegion RPC to an RS, we are catching NSRE but when the RS is the one throwing the NSRE, then it comes back as a RemoteException (then an NSRE) and we aren't unwrapping it properly.
> We need to catch this and then deal with it appropriately.
> Still tracking how this happened in the first place.

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

[jira] Assigned: (HBASE-3163) If we timeout PENDING_CLOSE and send another closeRegion RPC, need to handle NSRE from RS (comes as a RemoteException)

Posted by "Jonathan Gray (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/HBASE-3163?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jonathan Gray reassigned HBASE-3163:
------------------------------------

    Assignee: Jonathan Gray

> If we timeout PENDING_CLOSE and send another closeRegion RPC, need to handle NSRE from RS (comes as a RemoteException)
> ----------------------------------------------------------------------------------------------------------------------
>
>                 Key: HBASE-3163
>                 URL: https://issues.apache.org/jira/browse/HBASE-3163
>             Project: HBase
>          Issue Type: Bug
>          Components: master
>    Affects Versions: 0.90.0
>            Reporter: Jonathan Gray
>            Assignee: Jonathan Gray
>            Priority: Critical
>             Fix For: 0.90.0
>
>
> When we send a closeRegion RPC to an RS, we are catching NSRE but when the RS is the one throwing the NSRE, then it comes back as a RemoteException (then an NSRE) and we aren't unwrapping it properly.
> We need to catch this and then deal with it appropriately.
> Still tracking how this happened in the first place.

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

[jira] Commented: (HBASE-3163) If we timeout PENDING_CLOSE and send another closeRegion RPC, need to handle NSRE from RS (comes as a RemoteException)

Posted by "Jonathan Gray (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HBASE-3163?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12926612#action_12926612 ] 

Jonathan Gray commented on HBASE-3163:
--------------------------------------

And I think if we're going to do a sweep up of shit, we should just kill root.  It's going to make life so much easier and will be 90% ripping out code.

> If we timeout PENDING_CLOSE and send another closeRegion RPC, need to handle NSRE from RS (comes as a RemoteException)
> ----------------------------------------------------------------------------------------------------------------------
>
>                 Key: HBASE-3163
>                 URL: https://issues.apache.org/jira/browse/HBASE-3163
>             Project: HBase
>          Issue Type: Bug
>          Components: master
>    Affects Versions: 0.90.0
>            Reporter: Jonathan Gray
>            Assignee: Jonathan Gray
>            Priority: Critical
>             Fix For: 0.90.0
>
>
> When we send a closeRegion RPC to an RS, we are catching NSRE but when the RS is the one throwing the NSRE, then it comes back as a RemoteException (then an NSRE) and we aren't unwrapping it properly.
> We need to catch this and then deal with it appropriately.
> Still tracking how this happened in the first place.

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

[jira] Commented: (HBASE-3163) If we timeout PENDING_CLOSE and send another closeRegion RPC, need to handle NSRE from RS (comes as a RemoteException)

Posted by "HBase Review Board (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/HBASE-3163?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12926528#action_12926528 ] 

HBase Review Board commented on HBASE-3163:
-------------------------------------------

Message from: stack@duboce.net

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://review.cloudera.org/r/1130/#review1719
-----------------------------------------------------------

Ship it!


+1 for now.  I think yeah that we are going to have to do a sweep of these handlers and the Meta* classes and make them all do the same exception handling catching the same types but letting out (and aborting) for others -- but lets get a bit more experience first.

- stack





> If we timeout PENDING_CLOSE and send another closeRegion RPC, need to handle NSRE from RS (comes as a RemoteException)
> ----------------------------------------------------------------------------------------------------------------------
>
>                 Key: HBASE-3163
>                 URL: https://issues.apache.org/jira/browse/HBASE-3163
>             Project: HBase
>          Issue Type: Bug
>          Components: master
>    Affects Versions: 0.90.0
>            Reporter: Jonathan Gray
>            Assignee: Jonathan Gray
>            Priority: Critical
>             Fix For: 0.90.0
>
>
> When we send a closeRegion RPC to an RS, we are catching NSRE but when the RS is the one throwing the NSRE, then it comes back as a RemoteException (then an NSRE) and we aren't unwrapping it properly.
> We need to catch this and then deal with it appropriately.
> Still tracking how this happened in the first place.

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