Page 1 of 1

Corrupt Mailbox?

Posted: Mon Jun 13, 2005 4:15 am
by rrohde
Hallo,

I got another Scalix Problem!

A rather important mailbox seems corrupt! Some Mails cannot be displayed anymore... Below's some output from 'ommon'. How can this be fixed?

Rainer



ERROR Browser (Mime_Browser ) 06.07.05 14:08:30
[OM 2203] Characters in the string must be in one of the the ranges 0-9, a-z.
Pid of logging process: 28273


ERROR Browser (Mime_Browser ) 06.07.05 14:08:30
[OM.MIME 4000] Browser Args :mime.browse -l C -o /var/opt/scalix/temp/mime_cache/mime7ae4Tv ~/user/g000039/0000001:1,-32763
Pid of logging process: 28273


SERIOUS ERROR Remote Client (HTML Access ) 06.07.05 14:09:27
[OM 10270] Process about to terminate due to error.
Signal (Segmentation Violation) trapped by process 21710
Procedure trace follows:
-> htuICE_ProcessRequest
-> htuICE_InitCmdArgList
<- htuICE_InitCmdArgList
-> htuICE_GetConfig
<- htuICE_GetConfig
-> htuICE_RecvRequest
-> htuICE_ReturnHTML
-> htuICE_GetForm
<- htuICE_GetForm
-> htuICE_GetForm
<- htuICE_GetForm
-> htuICE_ProcessMarkups
-> htuICE_ReturnHTMLMarkup
<- htuICE_ReturnHTMLMarkup
-> htuICE_ReturnHTMLMarkup
-> htuICE_GetMarkupString
Pid of logging process: 21710


--------------------------------------------------------------------

Scalix Daemons' Stderr/Stdout


Ctner Access Monitor: A Dis-Interest request has been received which would
corrupt the access counts if serviced.
Pid = 27417
Filename = ~/data/000007q/003fus2
Ctner Access Monitor: A Dis-Interest request has been received which would
corrupt the access counts if serviced.
Pid = 27417
Filename = ~/data/00000af/003nhs8
Ctner Access Monitor: A Dis-Interest request has been received which would
corrupt the access counts if serviced.
Pid = 27417
Filename = ~/data/00000b1/003rvbh

--------------------------------------------------------------------

Scalix's/ Error Queue Contents



-------------------

Reference Number: 7662496
Ack Id: H0000091002e3545.1113202631.titan.ecommit.de
Sender:
Subject:
Message Type: Acknowledgement report
Sent: 04.11.05

-------------------

Reference Number: 8498634
Ack Id: H00000810032772f.1115039526.titan.ecommit.de
Sender:
Subject:
Message Type: Acknowledgement report
Sent:

-------------------

Reference Number: 9327546
Ack Id: H0000081003727d7.1117004299.titan.ecommit.de
Sender:
Subject:
Message Type: Acknowledgement report
Sent:

--------------------------------------------------------------------

Aborted Scalix Services


Omscan Server

Posted: Tue Jun 14, 2005 5:53 am
by ScalixSupport
It is possible there are two problems here.

For the first, do you know if this mailbox contains folders with a very large number of items in it, e.g. greater than approx, 32,000 items?

If this is the case, then the immediate short term solution will be to try and remove or copy out some of these items to another folder.

For the second, which doesn't look related (different pid), it would appear that the Server Rules Wizard was being used and hit some problem. Do you know or can you re-produce what operation was being done at the time.

Cheers,
JG

Posted: Tue Jun 14, 2005 11:56 am
by rrohde
Ok! I will try and get more precise info on this!

Posted: Mon Jun 20, 2005 2:34 am
by rrohde
I have gotten more information here, but the customer is German, so here is his reply in German!

I hope that one of the German speaking Scalix Support Reps can help us out here

Vielen Dank für Ihre Antwort auf meine Anfrage !

>For the first, do you know if this mailbox contains folders with a very
>large number of items in it, e.g. greater than approx, 32,000 items?

Ja, die Mailbox enthält mit Sicherheit mehr als 32000 items... allein 28000 Mails in der Inbox !!!
Gibt es ein bekanntes Limit, ab dem der Scalix Probleme bekommt ???



>If this is the case, then the immediate short term solution will be to
>try and remove or copy out some of these items to another folder.

Das heißt Scalix mag keine großen Ordner, wo liegen die Grenzen ?



>For the second, which doesn't look related (different pid), it would
>appear that the Server Rules Wizard was being used and hit some problem.
>Do you know or can you re-produce what operation was being done at the
>time."

Hmm, gute Frage...
Betrifft der zweite Punkt die gleiche Mailbox ?
Also am RulesWizard wurde in Bezug auf die problematische Mailbox nichts verändert !
Es lässt sich nicht reproduzieren... außerdem auch eine heikle Baustelle !!!

Posted: Tue Jun 21, 2005 12:18 am
by florian
Focusing on the first problem, yes, Scalix does have a limit on the maximum number of items PER FOLDER (size of items, size of mailbox in total is completely irrelevant). I believe it's a 2^15-1, i.e. 32767 limit, but JG might want to correct me here...

You will need to move items to other folders to resolve this.

(I'm impressed, btw - my Inbox is 7000 items and I thought that's large!)

Cheers,
Florian.

Posted: Fri Jun 24, 2005 10:57 am
by rrohde
Zu unserem letzten Problem:
Diese Meldungen des "System monitor" sind nun verschwunden:
- [OM 2203] Characters in the string must be in one of the the ranges
0-9, a-z.
- [OM.MIME 4000] Browser Args :mime.browse -l C -o
/var/opt/scalix/temp/mime_cache/mime1aHUar ~/user/g000039/0000001:1,-32762

Warum, weiß ich nicht, ich kann lediglich sagen, daß nach dem letzten
Wochenende die Meldungen weg waren !
Das einzige was ich bezügl. der Mailbox getan habe, ist sie am Freitag
vorher in eine Datei zu exportieren (omcpoutu).

Die nichtlesbaren Mails der Inbox waren immer noch nicht lesbar oder
vorhanden, aber konnten jetzt gelöscht werden.
Vorher hat ein Anklicken dieser Mails im Mailclient (Thunderbird) den
Mailclient in eine Schleife geschickt.

Komisch, noch ne Idee dazu ?

Posted: Fri Jun 24, 2005 1:18 pm
by ScalixSupport
Please can you post your questions in English. This will allow other non-German speaking forum members to follow the thread.

The error message that you are seeing is directly related to the fact that you have a very large number of messages in the folder.

The variable holding the item number is 16-bit and needs to be 32-bit. This is currently being reviewed and will be fixed in a later release.

In the short term, Florian's advice still stands that you should look to move a large number of messages out of the folder to get the message count below (approx) 32767.

Cheers

Dave.

Posted: Fri Jun 24, 2005 1:22 pm
by ScalixSupport
When a message is viewed using the IMAP client, it is passed through the MIME Browser process. This keeps a cache of rendered messages so that it does not need to generate them again (unless the message content has changed).

It's possible that you are coming up against a cached entry.

You can stop and restart the MIME browser process by running the command:

Code: Select all

omoff -d 0 mbc
omon mbc


When you have done that, log in with your IMAP client and attempt to access the message again.

Cheers

Dave

Posted: Thu Apr 19, 2007 2:47 pm
by souperdad
so if you have a couple of mailboxes that have become corrupted due to exceeding the folder size limits is it possible that this could in any way affect the server? IE could it spin out of control and use all of your available memory and swap?