db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Suresh Thalamati (JIRA)" <j...@apache.org>
Subject [jira] Commented: (DERBY-2107) Move page latching out of the lock manager
Date Fri, 02 Feb 2007 20:18:05 GMT

    [ https://issues.apache.org/jira/browse/DERBY-2107?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12469855
] 

Suresh Thalamati commented on DERBY-2107:
-----------------------------------------

Please commit the patch, you can address tests or comment fixes in 
separate patches. 

>2) I don't think the observer is used to clean up and release the
l> atch if the thread errors out, but rather to release the latch if the
>>container is closed/dropped while the page is latched. I

My understanding of this is  cleanup will tigger abort, or save point 
rollback which will trigger the close of container handle, which inturn 
will trigger release of latches on pages that are on the observer 
list of the container handle. 

I agree with you in most of the cases you will find latches being 
released in  the finally{ } blocks. I belive it is good practice too. But 
I am not sure that is case all  throughout the code, in any case i belive
cleanup will release the latches, if there are any of those missing cases. 

>> 3 and 5) I agree that the wait() should be timed. I was planning to
>change it in a followup patch if that sounds OK. What would be a
>reasonable timeout value? Ten seconds? 
>> 

I was thinking more like 1/2 hour or more,   mainly to get some kind of stack/information
 
and avoid the complaints that the derby  freezes forever without any clues, especially 
when derby is embedded in another application;   It may be the case that there are no 
bugs and will  not be any bugs in this area, so it may not be worth adding a timeout. 
I agree with Dan, this  can be addressed as separate issue, if needed. 
 
-suresh


> Move page latching out of the lock manager
> ------------------------------------------
>
>                 Key: DERBY-2107
>                 URL: https://issues.apache.org/jira/browse/DERBY-2107
>             Project: Derby
>          Issue Type: Improvement
>          Components: Performance, Services, Store
>    Affects Versions: 10.3.0.0
>            Reporter: Knut Anders Hatlen
>         Assigned To: Knut Anders Hatlen
>            Priority: Minor
>         Attachments: derby-2107-1a.diff, derby-2107-1a.stat, derby-2107-1b.diff, derby-2107-1c.diff,
derby-2107-1c.stat, derby-2107-1d.diff
>
>
> Latching of pages could be done more efficiently locally in store than in the lock manager.
See the discussion here: http://thread.gmane.org/gmane.comp.apache.db.derby.devel/33135

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


Mime
View raw message