[Dovecot] Connection to server time out when deleting (moving to Trash)

Dirk Essl de at floydworld.de
Fri Nov 24 17:41:17 UTC 2006


Anders Lund wrote:
> Dirk Essl wrote:
>> Anders Lund wrote:
>>> Hi,
>>>
>>> Running:
>>>  * Dovecot 1.0 rc15
>>>  * Debian Sarge Linux
>>>
>>> Using Maildir and IMAP over SSL.
>>>
>>> When trying to delete a lot of mail (> 5000-6000 messages) in some
>>> folder, using Mozilla Thunderbird (1.5.0.7) on Ubuntu, I often
>>> experience "Connection to server <hostname> timed out." Any suggestions
>>> on how to fix this problem?
>>>
>>> "Empty Trash" seem to work fine, so I guess this is related to moving a
>>> lot of mail from one folder to another (Thrash)?
>> I have exactly the same problem here. Solaris 10, dovecot 1.0rc2,
>> maildir, maildir space on a netapp filer mounted via NFS, Thunderbird 1.5.7
>>
>> I think in my case the filesystem is to slow. Could that be?
> 
> On my system I'm not mounting filesystem via NFS.
> 
> Using
> 
> Fibre Channel: LSI Logic / Symbios Logic FC919X Fibre Channel Adapter
> 
> attached directly to a
> 
> Vendor: NEXSAN    Model: ATAboy(9E263CD3)  Rev: A035
> 
> for the filesystem my mail is delivered to.
> 
> RAID5 on the ATAboy
> Generated ext3 filesystem using
>  * stride=16
>  * dir_index
> Mounting ext3 using data=writeback.

Can anyone pls test, if dovecot behaves the same on your system?
I think I'm not even able to delete 1000 messages in a row.

Timo: what would you need from us to help us fixing this?


> - Anders
> 
--
kind regards,
\dirk


More information about the dovecot mailing list