On 29-01-14 12:39, Reindl Harald wrote:
>
>
> Am 29.01.2014 11:26, schrieb Paul J Stevens:
>> On 29-01-14 10:42, Reindl Harald wrote:
>>> http://git.dbmail.eu/paul/dbmail/commit/?h=dbmail_3_1
>>>
>>> Revert "IMAP: defer bailout in case of EOF"
>>> This reverts commit afb1d1949e7bb9849e0fe33e4d3ae971711337bd
>>> __________________________________________
>>>
>>> is that something which should be done on prodcution servers
>>> independent of a future 3.1.11 release?
>>
>> Only if you suffer from a lot of connections in CLOSE_WAIT state:
>>
>> lsof -p `pidof dbmail-imapd`|grep CLOSE_WAIT
>
> not that much i would say currently
> the dovecot-proxy in front mitigates most likely
>
> may this partly explain the high memory corruption leading in
> imapd daily crashing after exceed 1.5 GB where i am still unsure
> if that maybe caused by libzdb and not dbmail itself
>
> [root@localhost:~]$ lsof -p `pidof dbmail-imapd`|grep CLOSE_WAIT | wc -l
> 535
I should think so. Those CLOSE_WAIT state connections refer to client
connections that are not closed at the server side, implying memory
bloat. So yes, CLOSE_WAIT connections in dbmail will cause leakage.
--
________________________________________________________________
Paul J Stevens pjstevns @ gmail, twitter, github, linkedin
* Premium Hosting Services and Web Application Consultancy *
www.nfg.nl/i...@nfg.nl/+31.85.877.99.97
________________________________________________________________
_______________________________________________
DBmail mailing list
DBmail@dbmail.org
http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail
