SWA slow for large mailbox

[natediggs]

Hi Everyone,

I've been using Scalix for a few months now. I just brought a new customer online who has a huge mailbox with over 4,000 contacts. While I understand that this is a really large mail box SWA runs VERY slowly when trying to add contacts to a new email. Is there any way to speed this up?

Looking in the scalix-swa.log I have several of these entries:
2009-07-12 23:02:02,034 WARN [SoapMail.fetchMessageHeaders:1420] Reverting to I
MAP to fetch headers because the platform failed with status 500 for request

Any help would be greatly appreciated.

Nate

[natediggs]

Now, when trying to search for a contact in the user's list of contacts I get an error that the connection has timed out. The search for contact dialog then says that an error has occurred and to contact the system administrator.

[les]

The most obvious question is what scalix version are you using?

There have been many many improvements (especially in swa and imap) in recent releases, 11.4.4 is the latest. Recommend you upgrade to 11.4.4, clear imap caches and see how things perform then.

[natediggs]

I am using 11.4.4.

[les]

I am using 11.4.4.

What about clearing the imap cache of the user?

[natediggs]

I have not done that and I will now. The user is a brand new user though. I created the account and uploaded the data on Friday.

[natediggs]

I just went to delete the user's IMAP cache and the user's account doesn't actually have an imap-cache folder.

[les]

I just went to delete the user's IMAP cache and the user's account doesn't actually have an imap-cache folder.

Highly unusual. It will get created if it doesn't exist. Sorry for asking.......but are you sure you're in the right user directory?

If so are permissions right (omcheck)? Does anything log to /var/opt/scalix/??/s/logs/fatal as you try to login?

[natediggs]

OK, I had the wrong user. There are two users with the same last name on the server and omshowu only showed one of them. My mistake.

I deleted the user's imap-cache folder and restarted IMAP. It rebuilt the cache.

When I log in as the user, go to create a new message, click to, and search for my name in his contact list it takes about a minute before it returns any results. It often times times out before it returns any results.

[natediggs]

In trying to trouble shoot this issue I looked at the fatal log and found several of these entries:


ERROR Indexer (Indexer ) Sun Jul 12 23:49:23 2009
[OM.DMON 3007] The SIS has been unable to process the Indexer's request. Details
:
SIS could not handle this request.
Reason: 500 (Internal Server Error)
Indexer Service has discarded the message:
User name: XXXXXXXXXXXXXXXXXXXXXXXXX
IndexId: 31e6b7d1-4a2184c3-4a579c2c-2b0c1
Attempted action: Add item 0001fa513a92928e to folder 0001aba61006be7c
Pid of logging process: 2091

[natediggs]

Anybody have any suggestions on speeding this up?

[les]

In trying to trouble shoot this issue I looked at the fatal log and found several of these entries:


ERROR Indexer (Indexer ) Sun Jul 12 23:49:23 2009
[OM.DMON 3007] The SIS has been unable to process the Indexer's request. Details
:
SIS could not handle this request.
Reason: 500 (Internal Server Error)
Indexer Service has discarded the message:
User name: XXXXXXXXXXXXXXXXXXXXXXXXX
IndexId: 31e6b7d1-4a2184c3-4a579c2c-2b0c1
Attempted action: Add item 0001fa513a92928e to folder 0001aba61006be7c
Pid of logging process: 2091

Certainly not normal to have sis errors in the fatal logs and could be the reason why things are slow.

there have been some recent posts talking about sis errors, search for something like "The SIS has been unable to process the Indexer's request" on the forum. Some of those recent posts may have a solution/workaround. if nothing helps you may have to open a support ticket directly with Scalix.

[RSisco]

natediggs,

If you do figure out what's going on, make sure to post an update in here.

Richard

[perbilse]

Hi,

we had the exact same problem. The cause was that the scalix-postgres module had not been initialized (supposed to happen automatically the first time "/etc/rc.d/init.d/scalix-postgres start" is done), leaving messages like the following in the installation log:

scalix-installer-20090719.log:2009-07-19 16:15:33,102 ERROR Operation failed: Failed to initialize the PostgreSQL database.

This in turn caused the index service to not function, resulting in the fallback to IMAP.

The problem with scalix-postgres was that our server was installed with en_GB as the default character set, while the scalix-postgres module wanted to use UTF8; this doesn't work, though I don't know why. What worked for us was to specify "--locale=LATIN1" in the database initialization script. Look for "Initialize the database" in etc/rc.d/init.d/scalix-postgres, our code is on line 198 and now reads:

$SU postgres -c "$PGENGINE/initdb --locale=LATIN1 --pgdata='$PGDATA' --auth='ident sameuser'" >> "$PGLOG" 2>&1 < /dev/null

Note: In spite of having weird letters in the alphabet of my mother tongue, I've never looked into character sets much. There may be gotchas I'm not aware of.

Best,

-- Per

[RSisco]

Well, my install log doesn't complain about not being able to initialize PostgreSQL and I have no problems searching in Outlook - so as far as I know, the indexing is working fine. I did find the following in my previous install log (not my latest version upgrade log file)...

Code: Select all

scalix-installer-20090623.log:2009-06-23 12:09:09,561 INFO Checking PostgreSQL Database encoding
scalix-installer-20090623.log:2009-06-23 12:09:10,197 INFO PostgreSQL Database encoding is not Unicode - need update
scalix-installer-20090623.log:2009-06-23 12:09:10,858 ERROR Operation failed: Pre-upgrade function use_unicode_pg_db_encoding failed:  Could not recreate PostgreSQL DB.

My latest install log actually makes no mention of PostgreSQL at all.

So I am not sure if this fix applies to me or not. I'll investigate further though.

Thanks.

Richard