You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@felix.apache.org by "Jamie goodyear (JIRA)" <ji...@apache.org> on 2010/02/24 15:30:27 UTC
[jira] Created: (FELIX-2130) [Karaf] Failover Clustering Mechanism
- Database shutdown / loss results in all Karaf instances becoming Active
on restart of that DB
[Karaf] Failover Clustering Mechanism - Database shutdown / loss results in all Karaf instances becoming Active on restart of that DB
-------------------------------------------------------------------------------------------------------------------------------------
Key: FELIX-2130
URL: https://issues.apache.org/jira/browse/FELIX-2130
Project: Felix
Issue Type: Bug
Components: Karaf
Environment: Oracle XE Database backend for Karaf JDBC Lock.
Reporter: Jamie goodyear
Two part issue.
1) Oracle isAlive() method needs to poll its connection to DB to ensure connection alive.
2) Main#lock() remove second break statement - in the event of DB service loss this breaks the lock out of normal spin lock routine.
I'll be supplying a patch for this issue.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Assigned: (FELIX-2130) [Karaf] Failover Clustering Mechanism
- Database shutdown / loss results in all Karaf instances becoming Active
on restart of that DB
Posted by "Guillaume Nodet (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/FELIX-2130?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Guillaume Nodet reassigned FELIX-2130:
--------------------------------------
Assignee: Guillaume Nodet (was: Chris Custine)
> [Karaf] Failover Clustering Mechanism - Database shutdown / loss results in all Karaf instances becoming Active on restart of that DB
> -------------------------------------------------------------------------------------------------------------------------------------
>
> Key: FELIX-2130
> URL: https://issues.apache.org/jira/browse/FELIX-2130
> Project: Felix
> Issue Type: Bug
> Components: Karaf
> Environment: Oracle XE Database backend for Karaf JDBC Lock.
> Reporter: Jamie goodyear
> Assignee: Guillaume Nodet
> Attachments: felix2130.patch, felix2130.txt
>
>
> Two part issue.
> 1) Oracle isAlive() method needs to poll its connection to DB to ensure connection alive.
> 2) Main#lock() remove second break statement - in the event of DB service loss this breaks the lock out of normal spin lock routine.
> I'll be supplying a patch for this issue.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Resolved: (FELIX-2130) [Karaf] Failover Clustering Mechanism
- Database shutdown / loss results in all Karaf instances becoming Active
on restart of that DB
Posted by "Guillaume Nodet (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/FELIX-2130?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Guillaume Nodet resolved FELIX-2130.
------------------------------------
Fix Version/s: karaf 1.6.0
Resolution: Fixed
I've applied the first patch as I don't think having a growing delay for the keep alive is a good idea. I would be ok with a different value if there is a need, but using a growing delay would mean that after some time, the delay for the keep alive could be one day, which does not make sense i think.
> [Karaf] Failover Clustering Mechanism - Database shutdown / loss results in all Karaf instances becoming Active on restart of that DB
> -------------------------------------------------------------------------------------------------------------------------------------
>
> Key: FELIX-2130
> URL: https://issues.apache.org/jira/browse/FELIX-2130
> Project: Felix
> Issue Type: Bug
> Components: Karaf
> Environment: Oracle XE Database backend for Karaf JDBC Lock.
> Reporter: Jamie goodyear
> Assignee: Guillaume Nodet
> Fix For: karaf 1.6.0
>
> Attachments: felix2130.patch, felix2130.txt
>
>
> Two part issue.
> 1) Oracle isAlive() method needs to poll its connection to DB to ensure connection alive.
> 2) Main#lock() remove second break statement - in the event of DB service loss this breaks the lock out of normal spin lock routine.
> I'll be supplying a patch for this issue.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (FELIX-2130) [Karaf] Failover Clustering Mechanism
- Database shutdown / loss results in all Karaf instances becoming Active
on restart of that DB
Posted by "Jamie goodyear (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/FELIX-2130?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jamie goodyear updated FELIX-2130:
----------------------------------
Attachment: felix2130.patch
File: felix2130.patch
This updated file contains the changes from the original patch file, plus an incremental sleep that resets when a successful lock occurs.
> [Karaf] Failover Clustering Mechanism - Database shutdown / loss results in all Karaf instances becoming Active on restart of that DB
> -------------------------------------------------------------------------------------------------------------------------------------
>
> Key: FELIX-2130
> URL: https://issues.apache.org/jira/browse/FELIX-2130
> Project: Felix
> Issue Type: Bug
> Components: Karaf
> Environment: Oracle XE Database backend for Karaf JDBC Lock.
> Reporter: Jamie goodyear
> Assignee: Chris Custine
> Attachments: felix2130.patch, felix2130.txt
>
>
> Two part issue.
> 1) Oracle isAlive() method needs to poll its connection to DB to ensure connection alive.
> 2) Main#lock() remove second break statement - in the event of DB service loss this breaks the lock out of normal spin lock routine.
> I'll be supplying a patch for this issue.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (FELIX-2130) [Karaf] Failover Clustering Mechanism
- Database shutdown / loss results in all Karaf instances becoming Active
on restart of that DB
Posted by "Jamie goodyear (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/FELIX-2130?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jamie goodyear updated FELIX-2130:
----------------------------------
Attachment: felix2130.txt
File: felix2130.txt
Contains fixes for the two issues outlined in this Jira entry.
> [Karaf] Failover Clustering Mechanism - Database shutdown / loss results in all Karaf instances becoming Active on restart of that DB
> -------------------------------------------------------------------------------------------------------------------------------------
>
> Key: FELIX-2130
> URL: https://issues.apache.org/jira/browse/FELIX-2130
> Project: Felix
> Issue Type: Bug
> Components: Karaf
> Environment: Oracle XE Database backend for Karaf JDBC Lock.
> Reporter: Jamie goodyear
> Attachments: felix2130.txt
>
>
> Two part issue.
> 1) Oracle isAlive() method needs to poll its connection to DB to ensure connection alive.
> 2) Main#lock() remove second break statement - in the event of DB service loss this breaks the lock out of normal spin lock routine.
> I'll be supplying a patch for this issue.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Assigned: (FELIX-2130) [Karaf] Failover Clustering Mechanism
- Database shutdown / loss results in all Karaf instances becoming Active
on restart of that DB
Posted by "Chris Custine (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/FELIX-2130?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Chris Custine reassigned FELIX-2130:
------------------------------------
Assignee: Chris Custine
> [Karaf] Failover Clustering Mechanism - Database shutdown / loss results in all Karaf instances becoming Active on restart of that DB
> -------------------------------------------------------------------------------------------------------------------------------------
>
> Key: FELIX-2130
> URL: https://issues.apache.org/jira/browse/FELIX-2130
> Project: Felix
> Issue Type: Bug
> Components: Karaf
> Environment: Oracle XE Database backend for Karaf JDBC Lock.
> Reporter: Jamie goodyear
> Assignee: Chris Custine
> Attachments: felix2130.txt
>
>
> Two part issue.
> 1) Oracle isAlive() method needs to poll its connection to DB to ensure connection alive.
> 2) Main#lock() remove second break statement - in the event of DB service loss this breaks the lock out of normal spin lock routine.
> I'll be supplying a patch for this issue.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (FELIX-2130) [Karaf] Failover Clustering
Mechanism - Database shutdown / loss results in all Karaf instances
becoming Active on restart of that DB
Posted by "Guillaume Nodet (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/FELIX-2130?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12862585#action_12862585 ]
Guillaume Nodet commented on FELIX-2130:
----------------------------------------
Committing to https://svn.apache.org/repos/asf/felix/trunk ...
M karaf/main/src/main/java/org/apache/felix/karaf/main/Main.java
M karaf/main/src/main/java/org/apache/felix/karaf/main/OracleJDBCLock.java
Committed r939585
> [Karaf] Failover Clustering Mechanism - Database shutdown / loss results in all Karaf instances becoming Active on restart of that DB
> -------------------------------------------------------------------------------------------------------------------------------------
>
> Key: FELIX-2130
> URL: https://issues.apache.org/jira/browse/FELIX-2130
> Project: Felix
> Issue Type: Bug
> Components: Karaf
> Environment: Oracle XE Database backend for Karaf JDBC Lock.
> Reporter: Jamie goodyear
> Assignee: Guillaume Nodet
> Fix For: karaf 1.6.0
>
> Attachments: felix2130.patch, felix2130.txt
>
>
> Two part issue.
> 1) Oracle isAlive() method needs to poll its connection to DB to ensure connection alive.
> 2) Main#lock() remove second break statement - in the event of DB service loss this breaks the lock out of normal spin lock routine.
> I'll be supplying a patch for this issue.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (FELIX-2130) [Karaf] Failover Clustering
Mechanism - Database shutdown / loss results in all Karaf instances
becoming Active on restart of that DB
Posted by "javier (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/FELIX-2130?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12843766#action_12843766 ]
javier commented on FELIX-2130:
-------------------------------
hi,
remove the break statement is ok, but the log will grow a lot with a DB error in the loop of the Main#lock(). it would be great a sleep or incremental sleep
javier
> [Karaf] Failover Clustering Mechanism - Database shutdown / loss results in all Karaf instances becoming Active on restart of that DB
> -------------------------------------------------------------------------------------------------------------------------------------
>
> Key: FELIX-2130
> URL: https://issues.apache.org/jira/browse/FELIX-2130
> Project: Felix
> Issue Type: Bug
> Components: Karaf
> Environment: Oracle XE Database backend for Karaf JDBC Lock.
> Reporter: Jamie goodyear
> Attachments: felix2130.txt
>
>
> Two part issue.
> 1) Oracle isAlive() method needs to poll its connection to DB to ensure connection alive.
> 2) Main#lock() remove second break statement - in the event of DB service loss this breaks the lock out of normal spin lock routine.
> I'll be supplying a patch for this issue.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (FELIX-2130) [Karaf] Failover Clustering
Mechanism - Database shutdown / loss results in all Karaf instances
becoming Active on restart of that DB
Posted by "Jamie goodyear (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/FELIX-2130?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12862352#action_12862352 ]
Jamie goodyear commented on FELIX-2130:
---------------------------------------
Hi,
In the event that the DB is lost already exists a "Thread.sleep(lockDelay);" within the Main#lock method -- we can extend this out to be a larger value or as you suggested make this an incremental delay. Will post back an updated patch, thanks for that catch.
Jamie
> [Karaf] Failover Clustering Mechanism - Database shutdown / loss results in all Karaf instances becoming Active on restart of that DB
> -------------------------------------------------------------------------------------------------------------------------------------
>
> Key: FELIX-2130
> URL: https://issues.apache.org/jira/browse/FELIX-2130
> Project: Felix
> Issue Type: Bug
> Components: Karaf
> Environment: Oracle XE Database backend for Karaf JDBC Lock.
> Reporter: Jamie goodyear
> Assignee: Chris Custine
> Attachments: felix2130.txt
>
>
> Two part issue.
> 1) Oracle isAlive() method needs to poll its connection to DB to ensure connection alive.
> 2) Main#lock() remove second break statement - in the event of DB service loss this breaks the lock out of normal spin lock routine.
> I'll be supplying a patch for this issue.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.