Imap cache issues

[florian]

it's a temporary socket. you may want to shut down your scalix server with omshut, then remove all those socket files remaining in the "temp" directory, then restart with omrc, they should get recreated.

If you believe permission errors may be affecting your operations, you may want to run the omcheck tool (see manpage for details)

if all that doesn't help, the issue may need investigation by a qualified support person - please see www.scalix.com/support for details

Florian.

[BigBirdy]

Wel I tried shutting down scalix (omshut) and then I deleted all the socket files in the temp folder, and restarted scalix. At first all services seemed fine..

root@penguin:~# omstat -a
PC Monitor Started NON-STOP 0
Directory Relay Server Started 09:05:08
Notification Server Started 09:05:08 0
Shared memory daemon Started NON-STOP
Notification Monitor Started NON-STOP
Session Monitor Started NON-STOP
Indexer Started NON-STOP
Stats Daemon Started NON-STOP
Container Access Monitor Started NON-STOP
Item Structure Server Stopped 06.23.08
Database Monitor Started 09:05:08
Licence Monitor Daemon Started NON-STOP
LDAP Daemon Started 09:05:08
Queue Manager Started NON-STOP
Item Delete Daemon Started NON-STOP
IMAP Server Daemon Started 09:05:08
SMTP Relay Started 09:05:08
Mime Browser Controller Started 09:05:08
Event Server Started 09:05:08

But after trying to connect with Thunderbird I got the error message again and can see that the Mime Browser is down again.

root@penguin:~# omstat -a
PC Monitor Started NON-STOP 0
Directory Relay Server Started 09:05:08
Notification Server Started 09:05:08 0
Shared memory daemon Started NON-STOP
Notification Monitor Started NON-STOP
Session Monitor Started NON-STOP
Indexer Started NON-STOP
Stats Daemon Started NON-STOP
Container Access Monitor Started NON-STOP
Item Structure Server Stopped 06.23.08
Database Monitor Started 09:05:08
Licence Monitor Daemon Started NON-STOP
LDAP Daemon Started 09:05:08
Queue Manager Started NON-STOP
Item Delete Daemon Started NON-STOP
IMAP Server Daemon Started 09:05:08
SMTP Relay Started 09:05:08
Mime Browser Controller Aborted 09:05:08
Event Server Started 09:05:08

[florian]

yup, so it seems that there is a problem message that T'Bird is trying to retrieve. The MIME Browser prepares the MIME version of a message internally stored in Scalix format. you should see some errors in omshowlog or the fatal log file related to the error. Not sure if that output helps identifying the message. if it does, it could possibly be deleted using omcontain, restoring service (actually I would like to have a copy of the message created using sxmboxexp, so that we can analyze if it's just corrupt or if it's a certain valid (maybe spam?) message that our system doesn't cope with too well!)

I really cannot walk you through this process in detail, you will find hints on all the steps above across the forum, and again, commercial support is also available should you require serious help.

Florian

[tdegouw]

The post here belonged to another thread.

[florian]

Well, you're using Scalix 11.4.0; the thread you're posting in is related to Mime Browser problems, while what you're experiencing seems - at best - to be related to IMAP cache corruption; this is supposed to be fixed in 11.4.1-U1. Please work with support to understand what is behind your issue.

Florian.

[tdegouw]

Well, you're using Scalix 11.4.0; the thread you're posting in is related to Mime Browser problems, while what you're experiencing seems - at best - to be related to IMAP cache corruption; this is supposed to be fixed in 11.4.1-U1. Please work with support to understand what is behind your issue.

Florian.

Woops, wrong thread.. i'll post it to the other thread.

[BigBirdy]

omshowlog does not show which, if any, messages may be the problem. Are there any other suggestions that might help identify the problem message?

omshowlog only shows many of these lines

ERROR Indexer (Indexer ) 08.28.08 13:45:32
[OM 24070] Debug message for Lab use :
mbcl_connect: Failed to connect to unix socket: /var/opt/scalix/pn/s/temp/mime_browse

[florian]

at least not a simple one, now we're basically in the realm of starting to strace processes to see what they are doing and where; obviously, if it's a single user that's affected by it, you could also consider moving messages around to different folders, seeing to isolate it. I assume you've done a full omscan on the user's mailbox and it's at least structurally consistent?

Florian.

[BigBirdy]

Well I ran the command suggested to export/identify the message and then removed it using Outlook. I also then ran the imap cache cleaning script from the wiki, and then restarted scalix. But the same problem persists, the mime browser is dying.

I ran this to extract the message and then removed using Outlook and then cleaned the Outlook trash to be sure

sxmboxexp -a 00018405803b0e32.exp --dref 00018405803b0e32 -F

Is there anything else I can do do diagnose and/or fix this problem?

Thanks

[florian]

well, it may also be the message after the last message successfully accessed or else.

I'll repeat what I said - I don't think we can walk through the work to identify what's going on here in the forum, at least not from our side. I'd suggest that, if you need serious help with it, you open a support case with Scalix Technical Support, http://www.scalix.com/support tells you how. This is a for-cost option, however, you'll be credited back the support incident if it is found that the issue is caused by a problem within our software.

Best,
Florian.

[mmas]

Hi,
I have upgraded to 11.4.3 and now imap cache it seems to work correctly.
I manage also old scalix servers and seems that imap-cache problem appear when I reboot.
Which is the correct sequence to close scalix, tomcat, postgres and reboot server ?

[mmas]

Hi,

curious but this morning I have rebooted a scalix server 11.4.3 (RHEL5) after month
of uptime and no swa users problem.

An user after 2 hours phone me and say he can't see new emails via swa.

Omscan do nothing.

Clear imap cache solve the problem.

I think these issues are reboot / service restart related.

[florian]

did you actually have to clear the IMAP cache or was it just enough to make sure that all IMAP processes were terminated and restarted.

we have identified a few further issues with regards to this area of the product; those appear rarely in 11.4.3, predominantly in heavily loaded multi-server setups, and we'll provide further improvements to this in subsequent releases.

Florian

[mmas]

Well, I use the clear imap cache script found on wiki page.

So it kill imap processes and clear imap cache. No tomcat restart needed.
Next time I'll try to use only clear imap.

I restart the server using reboot command. Someone suggest me to run omshut before reboot.
One thing I've notice is some errors in pg_log on close.

[florian]

if it was a clean shutdown/reboot it should actually be able to recover. obviously using a method that shuts down scalix properly - would be tomcat first, then postgres, then the scalix server core, is a little bit safer. in any case, the sw should be robust enough to recover from it.

for your further testing, make sure whether, when the problem appears, it can be recovered from by just killing the user's IMAP processes w/o clearing the on-disk cache files. I think the wiki script (which was contributed, so i don't have the specs) would actually have an option to do just that (-k?? please check)

Florian