Re: [cobalt-users] Qube2 Crashes without obvious cause

[Malcolm McLeary <mmcleary@xxxxxxx>]

Hi Mike,

Thanks for listening.

on 16/1/01 3:07 AM, Mike Vanecek at nospam99@xxxxxxxxxxxx wrote:

> One can monitor the imap activity within in Outlook as it scans the server. I
> understand the client's needs as I use it much the same way. If it were me, I
> would check the disk space issue (don't know how much the imap process needs,
> nor where it needs it, especially for temp storage during a session).

They had been getting disk space warnings messages,  but tends to be usual
because Cobalt never allocated enough space to / anyway.

[root /root]# df -m
Filesystem         MB-blocks    Used Available Capacity Mounted on
/dev/hda1                290     227       63     78%   /
/dev/hda3                193       5      188      3%   /var
/dev/hda4              18172     461    17711      3%   /home
[root /root]#

Similarly, user and group quotas are large and very empty.

> :>Since this problem has developed over time I have a feeling that it is
> :>related to the gradual accumulation of email.  Hence I have suggested to the
> :>customer that they create a bunch of local folders instead of IMAP folders
> :>and see if stability returns.
> 
> The difficulty with that suggestion is that it removes the reason for imap
> folders - i.e., email mobility.  Gradual accumulation of email --> more disk
> space requirements?

Yes I know, but at this point in time stability is more important than email
mobility.  It is a small office, each user has their own PC and none are
into remote access.

> :>Any fresh ideas (or the solution) would be gratefully accepted.
> 
> Turn telnet (sshd surely), and have yu-lan do several send/receives while you
> monitor disk space. Have yu-lan monitor the OL2K imap activity during the same
> time. Hope you can find it, I know it has been a thorn under the saddle for
> you.

Since I now have two identical machines (config and data) I'll try to get
the live version stable via whatever means and work on finding the real
problem off line.  When the Qube goes down, it takes about 20 minutes to
come back as it has to do a disk check.  The users are getting REALLY
annoyed with the down time (and rightly so).

It might also be an interesting exercise to copy the email folders over to a
Qube3 to see if it suffers the same fate.  If it doesn't, then a solution
might be to upgrade a 6 month old machine to a Qube3.

Cheers,  Malcolm