Session Hang on logout

[Tim King]

I am still having issues with sessions hanging in the Destroying state when I log out. The only way I have been able to resolve this is to reboot the Linux application server. Is there a known issue or fix for this?

[nepdev]

Same issue here. Happens every few days. I have posted detailed log files to OVD support after it happened. 

Pls let me know the status of that issue.

[nepdev]

It seems that every time this error happens, I get this type of line in the application server log: 

2012-12-18 16:26:43,828 [ERROR]: [1040] Weird, running session 1355844089WAdzp no longer exist

So OVD knows that there is something weird going on. 

it seems most likely this is an internal OVD bug, and not really because you did something wrong on your setup. If anyone knows a tip how to get around this, please post here.

[nepdev]

 

 

Does anyone of you have a virtual setup, i.e. are you running the OVD server(s) as VMs? 

I have them running as 4 VMs under XenServer 6, 2 application servers, 1 file server, 1 session manager. 

Also had the same issue when running them all in one single VM. 

 

I just wonder if the problem is with the virtualized setup - are you running directly on hardware? 

 

On Thursday, December 13, 2012 10:54:01 AM UTC+1, Shakti D Rathore wrote:

Same issue here , please

 

 

 

let me know how fix

[Simón Alonso Sánchez]

yes. i have a similar ovd environment than you... without any issue related to session hangout running the trunk branch under kvm/qemu.

1 linux application server

1 windows application server

1 session manager

1 web client

1 file server

1 gateway server

It could be useful if you could check at your "session hang on" when you close the session if there is any still profile mounted. Check inside the application server where the profile is mounted. Try to replicate the issue disabling the use of profiles at the session manager ovd web console. If you are using a windows application server try to disable it from the ovd farm  and try to replicate the issue without the windows application server...

2012/12/18 nepdev <pschulz...@gmail.com>

--
Un Saludo :

Simón Alonso Sánchez
simo...@gmail.com

[nepdev]

 It looks like the issue was that one has to use the EXACT version of Ubuntu / CentOS etc which is stated in the Ulteo support documents.

All my installations were done on Ubuntu 12.04, or on Centos 6.0.
I reinstalled the application servers on 10.04 - and in the last 2 days I had no issue any longer with 25 users (whereas before I had every half an hour a problem and I had to restart the application servers).
The errors seem related to CIFS but I could not say for sure.

So my conclusion is - these new versions ARE NOT SUPPORTED with Ulteo OVD out of the box.
You MUST use Ubuntu 10.0.4 (or CentOS 5.5) as stated in the installation support document.

Hope this helps anyone who had this issue.

[nepdev]

It seems also necessary to set session time-out to "none".

The session time-out anyway does not work how one would expect it to work (i.e. idle period), so just disable it.

[nepdev]

One more thing - make sure you have enough RAM, and especially enough HDD space on the application server. Even if you have a separate file server, you would need minimally 16 GB for 25 users on each app server, because Ulteo copies the entire profile from the file server to the app server when a session starts. 

If it runs out of HDD space, then you get the weird "password failed" box.