SWA BROKE? Your Scalix server is temporarily unavailable.

[kcwhited]

I have a problem where SWA quits working. I have had the problem since I installed Scalix, I am running Centos 5.3 and have just upgraded to scalix 11.4.4 this problem existed in 11.4.3 before the upgrade.
I have 15 users and I am the only one that uses SWA. I also have two pcs that are connected to my mailbox, which may lead to the question does scalix have an issue with multiple concurent logins to one mailbox?
I am not sure if it is a multiple login problem because it does seem to happen whether or not my outlook is connected to scalix.

The problem can occur while I am logged into scalix SWA or after it just sits for a few days. If I am logged in it just stops responding and says it cannot connect the scalix server which isn't an internet related problem as I am on the local network. If it stops responding I simply close IE or FF and try to log back in and I am presented with a messages that states "Your Scalix server is temporarily unavilable. Please try again later." The only way I can regain control of Scalix is to restart the scalix service, the outlook / scalix connector continues to function properly. I can log into SAC but I cannot log into SWA with any user until I restart scalix.

When I restart scalix I do get a message that states
Stopping Scalix services (scalix): ***************** WARNING **************************
/opt/scalix/bin/omshut : Some UAL Interface processes are still running
after waiting for 300 seconds for their termination
To wait for a longer period run 'omshut -t <timeout>'.

***************** WARNING **************************
[ OK ]
Starting Scalix services (scalix): [ OK ]

After that everything will work for a day or three then I have to kick the service again. As far as I can tell there isn't anything consistent in the mailbox that can cause this problem, nor would I expect it too.

Does anyone have any ideas how I might kill this gremlin.

Currently I am only using scalix for my personal mail, but the intention of using it was to test before purchasing the enterprise version for the office. I certainly can't be restarting Scalix every other day if I were to deploy it in production. I know there has to be something that needs tweaked because certainly nobody else seems to be having a problem like this.

Here is my environment:
Windows server 2k8 sp2 with VMWARE SERVER 2.0.1
2x dual core opteron 2.0ghz
host 8gb ram
The scalix vm runs Centos 5.3 32bit with all yum updates. (tried the updates after I had this problem) - no joy
VM has 1536MB of ram and 50GB drive.
SCALIX 11.4.4

[MDAFederal]

I believe that I too am having the same or similar issue. One thing that I noticed is that they upped the default Reserved memory for Tomcat to 1024 MB. With only 1500 MB reserved for the VM, you are most likely running into a memory issue. Try reconfiguring Tomcat to reserve only 512 MB and see if it stabilizes.

[kcwhited]

Hey I am having trouble finding information about the location of this setting can you help me with that ?

I wanted to add that this problem seems to affect only SWA, I can get into SAC when the problem occurs, - very odd, I dont think its the above mentioned memory setting but I am willing to try. It doesn't always happen while I am logged into scalix. for instance a few minutes ago I tried swa for the first time in two days and I can't get in.

Thanks
-Kevin

[MDAFederal]

You can update this in one of two ways.

1) the easiest way is to rerun the installer and reconfigure scalix components, then select Tomcat, and change the memory. This should restart the process withthe new values.

2) You can manually reconfigure in the /etc/opt/scalix-tomcat/scalix-tomcat.conf and change the "JAVA_OPT=" variable to something like the following.
JAVA_OPTS="-server -Djava.net.preferIPv4Stack=true -Xms512m -Xmx512m"

I would then run /etc/init.d/scalix-tomcat stop
followed by /etc/init.d/httpd stop, then start httpd and scalix-tomcat.

[kcwhited]

Thanks for the reply, sorry to say I dont think it has anything to do with Tomcat. the setting was 256, that said I went ahead and reset it to 512mb just for grins. However when I stopped then started scalix-tomcat and httpd I still cannot get into SWA it says "server is temporarly unavailable. I can however get into SAC. Its very odd as the outlook clients work fine, its seems to be something getting stuck with SWA and Scalix.
Whatever happens is happening when I am not using SWA, for instance the last two times this has happened I have been logged out of SWA, I get the error when attempting to check my mailbox. It has also happened while I have been in my mailbox, but since I have been having this problem I have stopped using the webmail application as much. last night I tried to reach my mailbox from a client's computer and ofcourse wasn't able to.

Here is what I just did incase anyone sees anything different this time.

I edited
/etc/opt/scalix-tomcat/scalix-tomcat.conf and changed

# Additional arguments to be passed to Java:
JAVA_OPTS="-server -Djava.net.preferIPv4Stack=true -Xms256m -Xmx256m"

to

# Additional arguments to be passed to Java:
JAVA_OPTS="-server -Djava.net.preferIPv4Stack=true -Xms512m -Xmx512m"

Then I
[root@scalix scalix-tomcat]# /etc/init.d/scalix-tomcat stop
Stopping Tomcat service (scalix): Using CATALINA_BASE: /var/opt/scalix/sx/tom cat
Using CATALINA_HOME: /opt/scalix-tomcat
Using CATALINA_TMPDIR: /var/opt/scalix/sx/tomcat/temp
Using JRE_HOME:
[ OK ]
[root@scalix scalix-tomcat]# /etc/init.d/httpd stop
Stopping httpd: [ OK ]
[root@scalix scalix-tomcat]# /etc/init.d/httpd start
Starting httpd: [ OK ]
[root@scalix scalix-tomcat]# /etc/init.d/scalix-tomcat start
Starting Tomcat service (scalix): Using CATALINA_BASE: /var/opt/scalix/sx/tom cat
Using CATALINA_HOME: /opt/scalix-tomcat
Using CATALINA_TMPDIR: /var/opt/scalix/sx/tomcat/temp
Using JRE_HOME:
[ OK ]

Then I checked SWA and it still didn't work so I did

[root@scalix scalix-tomcat]# /etc/init.d/scalix restart
Stopping Scalix services (scalix): ***************** WARNING **************************
/opt/scalix/bin/omshut : Some UAL Interface processes are still running
after waiting for 300 seconds for their termination
To wait for a longer period run 'omshut -t <timeout>'.

***************** WARNING **************************
[ OK ]
Starting Scalix services (scalix): [ OK ]

Then SWA worked again!
This is a pain! thanks for your input.

[MDAFederal]

If you run top at a command prompt, what do you see? Is java sitting on top? If you do an omshowu <username> and then do a ps -ef | grep <Unix ID> from the omshowu, do you have any hung imap41d processes sitting out there for your user?

[kcwhited]

running top:
Java bounces to the top of the list every now and then but doesn't stick there when it does go to the top it shows only around 1.9% cpu


[root@scalix /]# omshowu Kevin.Whited
Authentication ID: Kevin.Whited
Globally Unique ID: xxxxxxxxxxxxxxxxxxxxxxxx id removed
User Name : Kevin C. Whited /CN=Kevin Whited
MailNode : scalix
Internet Address : "Kevin Whited" xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx emails removed
System Login : 60536
Password : set
Admin Capabilities : YES
Mailbox Admin Capabilities : NO
Language : AMERICAN
Mail Account: Unlocked
Last Signon : 07.08.09 10:15:07
Receipt of mail : ENABLED
Service level : 0
Excluded from Tidyall : NO
Recovery Folder visible : NO
User Class : Full
SIS URL : sxidx://scalix.evltech.com/****id removed
[root@scalix /]#

Here is the process list for my username, not sure if this means its stuck? I dont think so because I see similar results for users known not to have used SCALIX SWA
[root@scalix /]# ps -ef |grep ************id removed
root 9055 7311 0 10:47 pts/1 00:00:00 grep ************id removed


This time the problem happened while I was in SWA, and my outlook was not connected, I was deleting email from my mailbox, then it stopped responding and gave me a timeout error... retry? click ok and it comes back up, eventually when I click on messages I can see in SWA is says "The message could not be loaded due to connection loss. Please try again later. " in the view pane


Thanks for sticking with me
-Kevin

[kcwhited]

I just wanted to add this as a poke for someone at Scalix, I understand I am currently not a paying customer, but the reason I installed Scalix for my personal mail domains is to Evaluate scalix as an alternative to our aging Exchange 2003 server I have 50 users on exchange and other than this stupid problem with SWAI have been impressed with its capability and performance VS Exchange, but I certainly can't be restarting the mail server daily, if the system were live I definately would have more than one person using SWA, and I am using Centos 5.3 which as far as i can tell shouldn't cause any issues for scalix, Since the paid version of scalix is just a license code from where I am I think its safe to assume that this problem would / could exist on a production server


Any help from the scalix team on this one? I realize it could be some weird dependancy issue but this problem has existed for me since I initially installed Scalix on what was then a fresh install of Centos 5.2 I have installed nothing else other than scalix on this machine, I am well aware of the problems that can come from package update dependencies so I do refrain from running updates unless I have to
-Kevin

[MDAFederal]

Kevin,

In your example you posted, it is the System login that you need to do the grep using,

So it would be ps -ef | grep 60536 as you are looking for system processes attached to that user.

I am a bit wary of SWA in 11.4.4 since something has changed and I cannot put my finger on it. 11.4.3 was solid and stable for me, and I have reverted my externally facing machine to 11.4.3 because I have had 0 (zero) issues with users using it.

Have you dug around in /var/opt/scalix/??/tomcat/logs for any info regaring the issue. I would specifically look at the scalix-swa.log, catalina.out, and scalix-api.log

[MDAFederal]

Kevin,

Any luck finding your issue? Have you checked to see if there are any IMAP processes hung for the user in question?

Any chance of running this not on a VM? Any chance of increasing the available memory of the VM?

Any help I can provide to get another user away from MS would be well worth it.

[kcwhited]

Sorry I haven't replied in a few days, thanks for checking up. The thing continues to lock up or whatever we want to call it. I haven't been able to find anything out of the ordinary, since I last posted I haven't used swa it seems to stay locked once it gets that way. I have been fighting a transition for one of our VM servers to XenServer from ESX. unrelated for sure but I can say that VMWARE is more robust, everything runs on vmware, but the capabilitys of the XenServer product are much higher but I have not had a lot of luck converting vms

Back to this project, I have tried adding more ram to the VM. I think I will try and move this vm to ESX instead of vmware server, I know I had some stability issues when we used to run some MySql vms on VMWARE SERVER. ESX is so many more times more reliable. maybe that will eliminate my gremlin. I will let you know.

Thanks
-Kevin

[kcwhited]

Ok,
I think I found the cause of the problem but I am not sure what I can do about it, it seems that the problem with SWA is a symptom of a problem with my windows mobile phone syncing via imap.

Has anyone else tried syncing a windows mobile phone with imap and scalix email server? I have disabled auto sync on my phone and the server hasn't had any glitches in days.
-Kevin

[jfk]

Same problem here - tried increasing the tomcat memory to 512Mb as per previous notes here, no impact. But after the "Your Scalix server is temporarily unavailable. Please try again later" popup, I also get a second popup: "An error has occurred. Contact your e-mail administrator if the problem persists. #A001". OK'ing this results in the "accessing mailbox" progress bar continuing.

omshowlog -l 9 shows a lot of lines like this:

REPORT Administration(ommodu ) 06.09.09 10:45:54
[OM 8060] User "postmaster /mail/CN=postmaster" modified successfully

but there is no-one logged in on sac - ???

Any pointers much appreciated.

[praveena]

Same problem here - tried increasing the tomcat memory to 512Mb as per previous notes here, no impact. But after the "Your Scalix server is temporarily unavailable. Please try again later" popup, I also get a second popup: "An error has occurred. Contact your e-mail administrator if the problem persists. #A001". OK'ing this results in the "accessing mailbox" progress bar continuing.

omshowlog -l 9 shows a lot of lines like this:

REPORT Administration(ommodu ) 06.09.09 10:45:54
[OM 8060] User "postmaster /mail/CN=postmaster" modified successfully

but there is no-one logged in on sac - ???

Any pointers much appreciated.

Hi.. Today i solved SAC login problem . Now i am not in front of my scalix server so tomorrow i will post the solution .

I think that will help .