Solaris

Hi Guys !

I have found this problem many times that the swap space (/tmp space more precisely )of my Server becomes full specially when i run a complete back up of the server .

once the /tmp space is full I am not able to run simple commands as i get error like :

fork no space

on checking df -hk , i could see that /tmp space which is also the swap space becomes full . on checking /tmp directory i could see there was nothing in that , thus i realized that this was due toh the long activity of backup done before and the server is not releasing this space .

In order to release this space i rebooted the system and the /tmp space was empty and probelm was fixed .

Can anybody suggest any other way of releasing the /tmp space ?

I am using Solaris 10 .

/tmp being full is likely only a side effect, not the root cause. Either there is a memory leak in your backup software or your swap is undersized.
Are the backup processes finished when the problem occurs ?

Earlier problem only used to occur after the backup was complete .

Currently i am not backing up the server but the /tmp is becoming full and when we are running the many processes on the server it use to throw the error like :

fork not enough space ..

the RAM size of my server is 8Gb and the swap space is 15 GB with this .

I guess it was due to less memory and more processes running at a time however now even i have killed several processes still /tmp space is not getting released to server .

Again, /tmp available space is a side effect, not the root cause.

What you describe is simply a system with not enough virtual memory (aka swap). Just add some swap or identify what processes are allocating that much memory.

Yes you are correct my server memory is not much , also i can t allocate more swap space .

As menioned by you i have already checked the processes which take more memory and have killed them ..

Now the question is even after killing many processes the swap space stilll shows the same , thus it is not getting released by itself . If i will reboot the server it releases this swap space .

Is there any other way to release this swap space then to reboot the server ?

Thanks in advance !!!

Processes using swap space is managed by the kernel. You have to take away the processes that use the space, not the other way around.
If you need the additional swap space and have available disk space, you can create a swap file or files.

Not much what ?
Please post "vmstat 2 2", "swap -l", "swap -s" and "prtdiag | grep Memory" output.
Why ? Adding a swap area is quite a simple operation.
What processes are these ? Please post ps or prstat output.
Post the commands you use to measure it.
Obviously, but that's pretty extreme. Sure, but you have to understand first why do processes exhaust all your virtual memory first. Killing them blindingly is not the right approach.