SWA - Request to server has time out

[tsanchez88]

Since upgrading our Scalix Webmail frontend to 11.4.4, we seem to get time out messages occasionally when moving to view items within Webmail. The time out appears to be random and occurs while viewing both mail and other items such as the calendar or contacts. There is no firewall between the Scalix DB and Frontend. Rolling back to 11.4.3 resolves the problem. The main Scalix server is running 11.4.4 and I tested using the SWA install that is directly on the main server and it does not seem to suffer the same problem. There are at times long pauses while loading messages but it does not time out.

I believe the way webmail loads was changed in 11.4.4 would this action be normal?


Popup message:

"Your request to the server has timed out. Keep trying?"

[BlankSheet]

We are having the same issue. Most of our testing was done on 11.4.3 and this issue never cropped up. Now we are getting ready to migrate this weekend and some test imports are receiving this error.

[gdzucker]

I am experiencing the same issue. I try restarting tomcat and, once this is done, SWA seems to work for the first few users that log in (though it is still slow). However, it dies again very quickly and the resource usage on the service itself is not high enough to explain the SWA time-outs.

Any help would be greatly appreciated.

[les]

i found that if i had too many clients open from the one pc i got the timeout messages also straight after i upgraded and tested to ensure all was ok.

I had thunderbird (imap) and outlook (scalix connect) open and then tried SWA i could login, but got timeouts. If i closed thunderbird and outlook the timeouts stopped and swa worked normally.

[MDAFederal]

I too am receiving a similar issue. I just upgraded to 11.4.4 last night and all seems fine.
Once we began to get a normal load on the system, the webmail interface will go out to lunch and hang. I have seen this before with issues of many IMAP connections open, but I am not seeing an alarming number of IMAP connections running.

Something appears to be failing with tomcat and/or IMAP, but am not seeing it in the logs.

Outlook users are not affected.

Please help,
Scott

[billb3]

I too have received this error message since upgrading to 11.4.4, but when I just tried to re-create...could not make it happen.

Has anyone opened a support ticket or a bug in bugzilla?

Edit: Got it to happen, and on IE7 & Firefox 3.0.11. Creating a bug in bugzilla now...

[billb3]

Just created this bug:

https://bugzilla.scalix.com/show_bug.cgi?id=20234

[MDAFederal]

I just lowered my memory reserve for Tomcat to 768 from 1024. My server has 10 GB of RAM in it, but I have had this issue before.

It appears that the default from 11.4.3 was 512 MB. I have since dropped my Tomcat RAM reserve to 512 to see if it is more stable.

This is on two machine that should be able to support the larger RAM numbers.

Any thoughts on this being the culprit?

[tsanchez88]

I don't believe the number of messages or Server load is an issue.

It does seem to occur when accessing a folder for the first time, so once its cached the problem may disappear. As for server load I realized the problem right after upgrading our servers and it was off hours with no users known to be logged in. Finally with the size, I've experienced this on a folder with 152 messages all listed at 16k in size. I had to click the the "OK" option to retry about 10 times before the folder finally loaded and then I was able to open/browse all messages in the folder just fine.

I don't see anything in the logs, I've tried watching "scalix-swa.log" but see zero activity while using the webmail. Does anyone know where else I should be checking.

[les]

I don't believe the number of messages or Server load is an issue.

It does seem to occur when accessing a folder for the first time, so once its cached the problem may disappear. As for server load I realized the problem right after upgrading our servers and it was off hours with no users known to be logged in. Finally with the size, I've experienced this on a folder with 152 messages all listed at 16k in size. I had to click the the "OK" option to retry about 10 times before the folder finally loaded and then I was able to open/browse all messages in the folder just fine.

I don't see anything in the logs, I've tried watching "scalix-swa.log" but see zero activity while using the webmail. Does anyone know where else I should be checking.

You may be onto something there. When it happened to me it was straight after an upgrade and i did clear all imap-caches for all users before attempting to login.

Then i got the error while having 3 clients open, but closing 2 of them off seemed to fix it. In actual fact it may have been the imap-cache being filled up that fixed it.

[tonysu]

I've been seeing this too but has turned out to be a non-critical error.

Apparently, since the HTTP protocol is essentially an asynchronously "disconnecting" protocol, the SWA user can simply click on another object (eg another message, or even another message and back again) to invoke another HTTP GET, and the http "conversation" continues (no login or re-authentication issues). Usually within a minute or two, the connection seems to be re-established (perhaps the busy Server finally returns a response?).

This is somewhat curious because this is a life-saver in 11.4, I seem to remember for whatever reason on 11.2 that SWA sessions would sometimes freeze implying some kind of non-HTTP continuous connection, and if the client forcibly terminated the frozen connection (eg Taskmanager or similar), that often precipitated a Server crash (requiring reboot).

I vastly prefer the User simply clicking a new object to continue a Server Timeout issue rather than having to reboot the Server.

Tony

[SidebandSamurai]

Hi everyone,

I have this problem quite regularly. What I do which usually resolves this issue is run this script:

imap-cache.sh

The code is right here for you to download:

Code: Select all

#!/bin/bash
if [ "$1" = "" ]
then
  echo "usage: $0 username"
  exit 1
fi
REALPATH=`/opt/scalix/bin/omrealpath "~/"`
OMSHOWU=`/opt/scalix/bin/omshowu -n "$1" -f 2>/dev/null| grep "User Folder"`
if [ $? -eq 0 ]
then
   FOLDER=`echo $OMSHOWU | awk -F/  '{print $2 "/" $3 }' `
   ID=`/opt/scalix/bin/omshowu -n "$1" | grep "System Login :" | awk -F: '{print $2}' `
  echo -n "Do you really want to remove the imap-cache for $1 (y/N)"
  read DOIT
  if [ "$DOIT" = "y" ]
  then
    echo "locking mail-account ..."
    /opt/scalix/bin/ommodu -o "$1" -K
    echo "killing IMAP-processes for user $1 (uid $ID) ..."
    pkill -u $ID in.imap41d
    echo "removing $REALPATH$FOLDER/imap-cache ..."
    rm -rf  $REALPATH$FOLDER/imap-cache
    echo -n "do you want to recreate the mailboxcache for $1 which can take a while (y/N)"                                                                                                                     
    read DOIT2
    if [ "$DOIT2" = "y" ]
    then
        echo "killing UAL-remote processes for user $1 (uid $ID) ..."
        pkill -u $ID ual.remote
        echo "removing $REALPATH$FOLDER/00000v8.ofs ..."
        rm -rf  $REALPATH$FOLDER/00000v8.ofs
        echo "recreating mailboxcache for $1...."
        /opt/scalix/bin/sxmbcprep -B -u "$1"
    fi
    omtidyu -B -M -u ' $1/sbgk'
    echo "unlocking mail-account ..."
    /opt/scalix/bin/ommodu -o "$1" -k
    echo "done!"
  fi
else
  echo "Can not find $1"
fi


I downloaded this script from the Scalix website, I don't remember exactly where. I believe it was provided by the Scalix Team to resolve these types of issues.

You execute the script as follows. In this example, the user account is "Cookie Monster"

Code: Select all

./imap-cache.sh "Cookie Monster"  <- include the quotes they are important

The script will then look for the account on the scalix server. Once the account has been found it will ask you the following question:

Code: Select all

Do you really want to remove the imap-cache for Cookie Monster (y/N)

Answer yes to remove the cache info.

The script will then proceed to remove the cache, and prompt you with the next question.

Code: Select all

Do you really want to remove the imap-cache for Cookie Monster (y/N)y
locking mail-account ...
ommodu: The user was modified successfully
killing IMAP-processes for user Cookie Monster (uid  60538) ...
removing /var/opt/scalix/gy/s/user/g00003a/imap-cache ...
do you want to recreate the mailboxcache for Cookie Monster which can take a while (y/N)y

Answer yes to this question.

A word of caution when answering this question. It can take the server a couple of hours or more for large accounts to rebuild the imap cache info for the user. Remember the account stays locked out during this time. The user will not be able to log in and view or respond to e-mail while the cache info is being rebuilt. Here is what the script looks like as it rebuilds Cookie Monster's account:

Code: Select all

killing UAL-remote processes for user Cookie Monster (uid  60538) ...
removing /var/opt/scalix/gy/s/user/g00003a/00000v8.ofs ...
recreating mailboxcache for Cookie Monster....
Building mailbox cache for: Cookie Monster /sbgk,us/CN=Cookie Monster
Mailbox cache directory:    /var/opt/scalix/gy/s/temp/mbc_prep/106

Inbox                            5069 msgs (190799Kb)
  Sent Calendar Entries           909 msgs (107808Kb)
Outbox                             18 msgs     (17Kb)
  CM-ECF                          282 msgs   (1021Kb)
  MyFax Inbox                     140 msgs  (14643Kb)

Mailbox cache built for: Cookie Monster /sbgk,us/CN=Cookie Monster
Mailbox cache directory: /var/opt/scalix/gy/s/temp/mbc_prep/106
Total cache data size:   3Gb

omtidyu : [OM 8013] The name/mailnode contains invalid characters

unlocking mail-account ...
ommodu: The user was modified successfully
done!



After the script has completed all its processes, I log into Cookie Monster's account through the Scalix web interface and there are no more errors. I have had this script resolve 100% of these issues when they occur.

Give this a try and Good Luck!

Sincerely,

Sideband Samurai

[easysoft]

Hi

I have just made a fresh install of 11.4.5 Community Edition.
I have created a test account on the system, which was fine for a few days. It has about 10 emails in the INBOX.

Now, logging in via webmail, it gets to the point where it says "Accessing Mailbox" and then hangs at that point.

Periodically I'll get a popup window saying :
Your request to the server has timed out. Keep trying?
or
Your Scalix server is temporarily unavailable. Please try again later.

If I restart Tomcat it works for a few more logins, then the problem restarts.

I have used both FF3.5/Linux and IE6/WinXP when getting this error.

Any ideas how to resolve this? Is there an ETA for Scalix 11.5 ?

Regards
Arif

[easysoft]

Apologies - this was a 'false alarm'.
Looking at /var/opt/scalix/bi/tomcat/logs/scalix-api.log I saw :

2009-09-07 16:34:30,485 FATAL [ImapDataCache.getSession:148] could not create database connection
org.hibernate.exception.GenericJDBCException: Cannot open connection

On checking, I found the scalix-postgresql service wasn't running

On starting it up everything seems fine now.

Apologies again,
Regards
Arif