You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Sylvain Lebresne (JIRA)" <ji...@apache.org> on 2012/09/18 16:03:07 UTC
[jira] [Created] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Sylvain Lebresne created CASSANDRA-4679:
-------------------------------------------
Summary: Fix binary protocol NEW_NODE event
Key: CASSANDRA-4679
URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
Project: Cassandra
Issue Type: Bug
Affects Versions: 1.2.0 beta 1
Reporter: Sylvain Lebresne
Assignee: Sylvain Lebresne
Priority: Minor
Fix For: 1.2.0
As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sylvain Lebresne updated CASSANDRA-4679:
----------------------------------------
Attachment: (was: 0002-Start-RPC-binary-protocol-before-gossip.txt)
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0003-Remove-hardcoded-initServer-from-AntiEntropyServiceTes.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Yuki Morishita (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13486132#comment-13486132 ]
Yuki Morishita commented on CASSANDRA-4679:
-------------------------------------------
Patch looks good, though AntiEntropyServiceStandardTest/AntiEntropyServiceCounterTest is failing after applying this patch.
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0001-4679.txt, 0002-Start-RPC-binary-protocol-before-gossip.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sylvain Lebresne updated CASSANDRA-4679:
----------------------------------------
Attachment: 0002-Start-RPC-binary-protocol-before-gossip.txt
0001-4679.txt
I've rebased the patch and attaching a second patch to start the binary protocol server before gossip. This seems to be working as expected but the fact that bootstrap is broken on trunk (due to CASSANDRA-4764) makes it a bit harder to test correctly.
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0
>
> Attachments: 0001-4679.txt, 0002-Start-RPC-binary-protocol-before-gossip.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13486709#comment-13486709 ]
Sylvain Lebresne commented on CASSANDRA-4679:
---------------------------------------------
Ok, you're right and I've updated the first patch to add back the same 'is initialized' check. I leave the third patch though since that hardcoded call is still useless (but the test pass even if we don't remove it).
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0001-4679.txt, 0002-Start-RPC-binary-protocol-before-gossip.txt, 0003-Remove-hardcoded-initServer-from-AntiEntropyServiceTes.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Brandon Williams (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13490710#comment-13490710 ]
Brandon Williams commented on CASSANDRA-4679:
---------------------------------------------
Sorry, edited my comment to clarify while you were posting yours :)
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0001-4679.txt, 0002-Start-RPC-binary-protocol-before-gossip.txt, 0003-Remove-hardcoded-initServer-from-AntiEntropyServiceTes.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sylvain Lebresne updated CASSANDRA-4679:
----------------------------------------
Attachment: 0002-Start-RPC-binary-protocol-before-gossip.txt
0001-4679.txt
I've also updated the first patch to send a TOPOLOGY_CHANGE event when a node has moved since that was missing and trivial to add.
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0001-4679.txt, 0002-Start-RPC-binary-protocol-before-gossip.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jonathan Ellis updated CASSANDRA-4679:
--------------------------------------
Reviewer: yukim
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0
>
> Attachments: 4679.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Resolved] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sylvain Lebresne resolved CASSANDRA-4679.
-----------------------------------------
Resolution: Fixed
Alright, I've reverted the 'start thrift before gossip part of this patch'. That has never been the most important part anyway. This should fix any regression at least as far as this ticket is concerned.
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0001-4679.txt, 0002-Start-RPC-binary-protocol-before-gossip.txt, 0003-Remove-hardcoded-initServer-from-AntiEntropyServiceTes.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sylvain Lebresne updated CASSANDRA-4679:
----------------------------------------
Attachment: 0002-Start-RPC-binary-protocol-before-gossip.txt
0001-4679.txt
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0001-4679.txt, 0002-Start-RPC-binary-protocol-before-gossip.txt, 0003-Remove-hardcoded-initServer-from-AntiEntropyServiceTes.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sylvain Lebresne updated CASSANDRA-4679:
----------------------------------------
Attachment: 0003-Remove-hardcoded-initServer-from-AntiEntropyServiceTes.txt
Interesting. This is due to the hardcoded call to StorageServer.initServer() in AntiEntropyServiceTestAbstract. But I have absolutely no clue what we have that call. In fact, removing that call (patch 3 attached) fixes the test. I'm not totally sure why the test was working previously, maybe the 2 patch of this ticket just changed the timing of the server initialization triggering that issue?
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0001-4679.txt, 0002-Start-RPC-binary-protocol-before-gossip.txt, 0003-Remove-hardcoded-initServer-from-AntiEntropyServiceTes.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sylvain Lebresne updated CASSANDRA-4679:
----------------------------------------
Attachment: 4679.txt
Attaching patch that adds a new listener interface for "higher level" events than the gossip ones.
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0
>
> Attachments: 4679.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13490708#comment-13490708 ]
Sylvain Lebresne commented on CASSANDRA-4679:
---------------------------------------------
As much as I'm happy to revert the part about starting the thrift/binary protocol server first, I'd prefer some precision about "breaks all kind of things". That is, does that break things only if people query the thrift interface before MS and gossip are up (which would make sense, but in real life doesn't seem like very likely to happen (which doesn't mean we shouldn't fix it btw)), or does it break things in other cases?
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0001-4679.txt, 0002-Start-RPC-binary-protocol-before-gossip.txt, 0003-Remove-hardcoded-initServer-from-AntiEntropyServiceTes.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13466802#comment-13466802 ]
Jonathan Ellis commented on CASSANDRA-4679:
-------------------------------------------
We can always change that ordering. Maybe we should.
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 4679.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Reopened] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Brandon Williams (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Brandon Williams reopened CASSANDRA-4679:
-----------------------------------------
Reopening because (at least sometimes) this allows thrift to start before MS and gossip, which breaks all kind of things.
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0001-4679.txt, 0002-Start-RPC-binary-protocol-before-gossip.txt, 0003-Remove-hardcoded-initServer-from-AntiEntropyServiceTes.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sylvain Lebresne updated CASSANDRA-4679:
----------------------------------------
Attachment: (was: 4679.txt)
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sylvain Lebresne updated CASSANDRA-4679:
----------------------------------------
Attachment: (was: 0002-Start-RPC-binary-protocol-before-gossip.txt)
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0001-4679.txt, 0002-Start-RPC-binary-protocol-before-gossip.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sylvain Lebresne updated CASSANDRA-4679:
----------------------------------------
Attachment: 4679.txt
I've rebased the patch and added a short paragraph to the protocol doc to explain that UP events are usually sent a few milliseconds before the binary protocol server is fully up (since gossip is started before the protocol server).
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 4679.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Yuki Morishita (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13486905#comment-13486905 ]
Yuki Morishita commented on CASSANDRA-4679:
-------------------------------------------
+1 for all patches.
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0001-4679.txt, 0002-Start-RPC-binary-protocol-before-gossip.txt, 0003-Remove-hardcoded-initServer-from-AntiEntropyServiceTes.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sylvain Lebresne updated CASSANDRA-4679:
----------------------------------------
Fix Version/s: (was: 1.2.0)
1.2.0 beta 2
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 4679.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sylvain Lebresne updated CASSANDRA-4679:
----------------------------------------
Attachment: (was: 0001-4679.txt)
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0003-Remove-hardcoded-initServer-from-AntiEntropyServiceTes.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Comment Edited] (CASSANDRA-4679) Fix binary protocol
NEW_NODE event
Posted by "Brandon Williams (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13490699#comment-13490699 ]
Brandon Williams edited comment on CASSANDRA-4679 at 11/5/12 4:11 PM:
----------------------------------------------------------------------
Reopening because (at least sometimes) this allows thrift to start before MS and gossip, which is wrong, but also other things seem to be breaking like bootstrap.
was (Author: brandon.williams):
Reopening because (at least sometimes) this allows thrift to start before MS and gossip, which breaks all kind of things.
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0001-4679.txt, 0002-Start-RPC-binary-protocol-before-gossip.txt, 0003-Remove-hardcoded-initServer-from-AntiEntropyServiceTes.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sylvain Lebresne updated CASSANDRA-4679:
----------------------------------------
Attachment: (was: 4679.txt)
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13466824#comment-13466824 ]
Sylvain Lebresne commented on CASSANDRA-4679:
---------------------------------------------
I suppose we can, yes. Though I suspect there may be races so that some queries (I'm thinking schema altering ones though honestly I haven't checked) might trigger assertion errors or related if gossip isn't started. Probably worth checking though. Maybe in a separate ticket however since it's not directly related (clients will have a to handle the case where after a NEW_NODE the node cannot be joined anyway).
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 4679.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sylvain Lebresne updated CASSANDRA-4679:
----------------------------------------
Attachment: (was: 0001-4679.txt)
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0001-4679.txt, 0002-Start-RPC-binary-protocol-before-gossip.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CASSANDRA-4679) Fix binary protocol NEW_NODE
event
Posted by "Yuki Morishita (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-4679?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13486354#comment-13486354 ]
Yuki Morishita commented on CASSANDRA-4679:
-------------------------------------------
I found the difference between patched and trunk.
Your initServer tries to "join ring" even after server is initialized, whilst in trunk it is guarded by "initialized" check.
I think it is better to check if initialized before calling matbeJoinRing in your patch.
> Fix binary protocol NEW_NODE event
> ----------------------------------
>
> Key: CASSANDRA-4679
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4679
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.0 beta 1
> Reporter: Sylvain Lebresne
> Assignee: Sylvain Lebresne
> Priority: Minor
> Fix For: 1.2.0 beta 2
>
> Attachments: 0001-4679.txt, 0002-Start-RPC-binary-protocol-before-gossip.txt, 0003-Remove-hardcoded-initServer-from-AntiEntropyServiceTes.txt
>
>
> As discussed on CASSANDRA-4480, the NEW_NODE/REMOVED_NODE of the binary protocol are not correctly fired (NEW_NODE is fired on node UP basically). This ticket is to fix that.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira