Client session hangs server
Hi everybody...
We have a farm of 6 Sun Ray Server 200 Sun Ray clients
V20z Solaris 10x86 6/06 five of them
V20z Solaris 10x86 11/06 one of them (just to check have last patches and updates)
Suddenly clients connected to one Sun Ray Server hangs. The server answer pings but don't let login or if you are logged you can't type commands (slow, slow server)and you must reboot. Meanwhile the server boots the problem goes to another server... you must reboot. So now 2 servers are off line and you can imagine the slow scenario.
We have detected that the problem begins with this process...
PID USERNAME SIZERSS STATE PRI NICETIME CPU PROCESS/NLWP
5663 bib421212M 1178M sleep5900:00:03 2.8% metacity/1
237 root322M 214M sleep5900:28:54 0.5% java/116
metacity process is more larger than others....
In the Sun Ray client where bib42 is working, in the task bar you can see Geko's logo (normally you don't see this item).
If you don't kill the process the server slows... until it hangs.
Sometimes the same account creates the same problem in other servers... so we tried to delete .metacity and the problem finish. But others accounts generate that process and it's an endless history.
Have you any information about problems with metacity?
I asked in this forum about install containers to isolate application but Otto answer that it was not the solution.
How can I limit the application.
The actual solution is delete all .files in each home directory and restart all servers in night. But if you don't restart servers 2 or 3 days the problem continues.
The servers have the last update.
We'll appreciate your help ... because we have a lot of students and teachers that hated us.... and we want to go home early.... long nights trying to understand what happen.
