You are viewing a plain text version of this content. The canonical link for it is here.
Posted to derby-dev@db.apache.org by "Knut Anders Hatlen (JIRA)" <ji...@apache.org> on 2007/02/05 14:09:05 UTC

[jira] Created: (DERBY-2291) Add timeout on latch wait

Add timeout on latch wait
-------------------------

                 Key: DERBY-2291
                 URL: https://issues.apache.org/jira/browse/DERBY-2291
             Project: Derby
          Issue Type: Improvement
          Components: Store
    Affects Versions: 10.3.0.0
            Reporter: Knut Anders Hatlen
            Priority: Minor


If a bug causes a dead lock in the latching, Derby may hang infinitely. A timeout could make it easier to detect the cause of the hang.

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


[jira] Closed: (DERBY-2291) Add timeout on latch wait

Posted by "Knut Anders Hatlen (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/DERBY-2291?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Knut Anders Hatlen closed DERBY-2291.
-------------------------------------

    Resolution: Won't Fix

Thanks for all the input. I don't feel very strongly for a timeout either, so I'm closing the issue as "Won't Fix".

> Add timeout on latch wait
> -------------------------
>
>                 Key: DERBY-2291
>                 URL: https://issues.apache.org/jira/browse/DERBY-2291
>             Project: Derby
>          Issue Type: Improvement
>          Components: Store
>    Affects Versions: 10.3.0.0
>            Reporter: Knut Anders Hatlen
>            Priority: Minor
>
> If a bug causes a dead lock in the latching, Derby may hang infinitely. A timeout could make it easier to detect the cause of the hang.

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


[jira] Commented: (DERBY-2291) Add timeout on latch wait

Posted by "Bryan Pendleton (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/DERBY-2291?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12470244 ] 

Bryan Pendleton commented on DERBY-2291:
----------------------------------------

I'm not sure I totally agree. I think you could argue that just staying hung is better, in the hopes that eventually somebody will come along and grab a stack trace or attach a debugger or something like that, and then the dead lock will be directly apparent from the stack traces of the various threads. If you time out and break the deadlock, the information about the various thread stacks will be lost.


> Add timeout on latch wait
> -------------------------
>
>                 Key: DERBY-2291
>                 URL: https://issues.apache.org/jira/browse/DERBY-2291
>             Project: Derby
>          Issue Type: Improvement
>          Components: Store
>    Affects Versions: 10.3.0.0
>            Reporter: Knut Anders Hatlen
>            Priority: Minor
>
> If a bug causes a dead lock in the latching, Derby may hang infinitely. A timeout could make it easier to detect the cause of the hang.

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


[jira] Commented: (DERBY-2291) Add timeout on latch wait

Posted by "Daniel John Debrunner (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/DERBY-2291?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12470252 ] 

Daniel John Debrunner commented on DERBY-2291:
----------------------------------------------

>From a quick search there are around thirty places in Derby that use Thread.wait(). Should there be timeouts on all of those as well? Or what makes latching unique that a timeout would be required?

I assume that a simple synchronized wait that's contained within a single object would not need a timeout since one could believe that code inspection and testing could demonstrate no chance for deadlock.

With latching now localized to the page object I assume we must be closer to a simpler model than before, I wonder if time would be better spent analyzing the code to prove deadlocks cannot occur or modifiying any api to remove the chance of an incorrect calling sequence that could lead to a deadlock.

> Add timeout on latch wait
> -------------------------
>
>                 Key: DERBY-2291
>                 URL: https://issues.apache.org/jira/browse/DERBY-2291
>             Project: Derby
>          Issue Type: Improvement
>          Components: Store
>    Affects Versions: 10.3.0.0
>            Reporter: Knut Anders Hatlen
>            Priority: Minor
>
> If a bug causes a dead lock in the latching, Derby may hang infinitely. A timeout could make it easier to detect the cause of the hang.

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