Solstice Snmp subagent anormal stop
Hi,
My subagent is stopping anormaly after a few hours at the reception of an Unknown signal 15/16 or 17 ( respectively SIGTERM, SIGUSR1 and SIGUSR2).
The call of the agent_end function is then done and the agent exits.
Is there someone who know why a such signal is sent to the subagent ?
Is it coming from the master agent ?
Will the apply of the 107709-20 patch when the solaris 7 computer was booted in 64bits might be the cause of the trouble ( this issue is appearing in 32 b boot mode, it has no be detected in 64b) ?
Thanks for your support,
christophe.
[620 byte] By [
feuillec] at [2007-11-26 5:54:05]
# 1
Updating the issue,
In fact when the agent_end function is called, it is not coming from an signal as described in developper guide ( the backtrace from gdb is different), the agent_end is called by error-exit which is called by snmp_loop....
This is occuring when the host is receiving bunches of snmp requests...
Is there a security which stops the agent when there is too much snmp requests onto ?
If it is, what are the limits ?
Is the poll-interval setting in snmpdx.rsrc can be the reason ?
Christophe.
>
> Hi,
>
> My subagent is stopping anormaly after a few hours at
> the reception of an Unknown signal 15/16 or 17 (
> respectively SIGTERM, SIGUSR1 and SIGUSR2).
>
> The call of the agent_end function is then done and
> the agent exits.
>
> Is there someone who know why a such signal is sent to
> the subagent ?
> Is it coming from the master agent ?
>
> Will the apply of the 107709-20 patch when the solaris
> 7 computer was booted in 64bits might be the cause of
> the trouble ( this issue is appearing in 32 b boot
> mode, it has no be detected in 64b) ?
>
>
> Thanks for your support,
>
> christophe.
>
