Forum Discussion
NimbostratusIs there a way (command etc.) to completely kill a specific process?
I want to test on the assumption that certain processes (bigd, cbrd, mcpd, scriptd, snmpd, sod, tmm, tmrouted, wccpd) have dropped on failure. (I want to capture what kind of syslog is output at this time)
Even if you kill a specific process with the kill command, the same process is restored immediately with another PID.
An example
ps aux | grep bigd
root 6456 0.0 0.0 1880 388? S 10:14 0: 00 runsv bigd
root 6479 0.0 0.3 43876 25240? S 10: 14 0: 02 bigd -S
root 19354 0.0 0.0 5568 828 ttyS 0 S + 11: 22 0: 00 grep bigd
kill 9 6456
kill 9 6479
-bash: kill: (6479) - No such process
The same process is restored even if you kill 6479 first.
ps aux | grep bigd
root 19392 0.0 0.0 1880 392? S 11: 24 0: 00 runsv bigd
root 19393 4.0 0.2 43736 24076? S 11:24 0:00 bigd -S
root 19469 0.0 0.0 5564 820 ttyS 0 S + 11: 24 0:00 grep bigd
Is there a way (command etc.) to completely kill a specific process?
2 Replies
- Kai_Wilke
MVP
Hi Ahan,
the
command is the right choice to terminate a given process. The problem in your case is just, that BIG-IPs watchdog service is automatically restarting thekill
service when it gets (for whatever reason) terminated. In addition to that you will see some lines in yourbigd
indicating the process restart .../var/logs/ltmTo stop (aka. no termination) the
service manually you may enter the following command on your bash...bigd~ bigstart disable bigdCheers, Kai
- boneyard
what are you trying to say Ahan? it seems to work now right? the second ps aux shows the /usr/bin/wccpd gone.
