System Lock table full

[Lester Grovers_1]

Hi

OVO reports the following error message "Can't lock queue file /var/opt/OV/tmp/OpC/actagtq (qd = 7, fd = ) due to full system table. No Locks available.
One of the instructions in the error message is to check the system lock table limits. Without a tool like Glance, how do i do this ?
Another instruction says that should qd or fd have values other than low positive interger numbers then this could possible be an internal error. Beside for the the values for FD being passed in the error message , how can i view what these are actually set to.

regards
Lester

[T G Manikandan]

The problem should be with the kernel parameter nflocks.

by default the value is 200.
you need to increase this.

check the kernel parameter page

http://docs.hp.com/hpux/onlinedocs/939/KCParms/KCparam.Nflocks.html

[Steve Steel]

Hi

read about the kernel parameters at www.docs.hp.com


In this case increase nflocks

Then you can lock more files.

sam
kernel config
configurable parameters
change nflocks
reboot


steve Steel

[Lester Grovers_1]

Hi ,

I am sorry, i should have previously mentioned in my original message that i have previously increase nflocks to a value of 12000.

regards
Lester

[Dietmar Konermann]

See:
http://www4.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000062906404

Best regards...
Dietmar.

[Steve Steel]

Hi


Try

Can't lock queue /var/opt/OV/tmp/OpC/actagtq (qd = 7, fd = ) due to full system table. No locks available.
qd is ok as a small integer, but fd isn't a small integer so this may be an internal problem.
As such stop the OpC agent
processes on the concerned system and then remove the queue files (deleting all files in /var/opt/OV/tmp/OpC/actagtq/ ).
Restarting the opc agents will recreate new virgin queue files.

To be sure take a tar of the QUEUE files before removing them. remove the tar later.

Steve Steel

[Lester Grovers_1]

Hi Steve,

Thanks for the Info, I can understand why stopping agent processes and removing q files could fix the problem. We have however had this problem before and have done exactly that. Why does this problem occur? I cannot see that inreasing the nflocks again as a valid option.

regards
Lester