[Mimedefang] mimedefang fails to start upon reboot

[jef moskot]

On Tue, 10 Nov 2009, Paul Murphy wrote:

>> Nov  7 23:45:30 phantom mimedefang-multiplexor[936]: Starting slave 1 (pid  1238) (1 running): Bringing slaves up to minSlaves (2)
>> Nov  7 23:45:30 phantom mimedefang-multiplexor[936]: Slave 1 stderr: Out of memory during request for 4088 bytes, total sbrk() is 9553920 bytes!
>> Nov  7 23:45:30 phantom mimedefang-multiplexor[936]: Slave 1 stderr: Out of memory during request for 48 bytes, total sbrk() is 9553920 bytes!
>> Nov  7 23:45:30 phantom mimedefang-multiplexor[936]: Reap: slave 1 (pid 1238) exited due to signal 11 (SLAVE DIED UNEXPECTEDLY)
>
>> To recap, upon reboot, mimedefang continually pukes out error messages
>> like the ones listed above until the process is manually stopped and
>> restarted.  After that, it works just fine.
>
> So the system runs OK under your environment, but not under the startup 
> environment.  From the error messages, I'd suspect that you have somehow 
> limited the memory use of processes on the system by default, and then 
> in your interactive environment you specifically remove any limits, so 
> when you start it manually all is OK.  To test this, include some code 
> in the system startup script to dump the current shell limits to a temp 
> file, and see what happens when you next restart the system.

Could you give a poor ignorant soul a hint as to how to go about that? 
I'm not sure what exactly I'm looking for.  Where can I find these 
variables in the current running config?

I didn't do any of that on purpose and the problem appeared at the same 
time as a perl upgrade, so I'd guess it was related to a change in either 
perl itself or the port.

One thing I notice is that there's a new perl.core in the defang queue 
each time I reboot.  Can anything useful be learned from that?

Jeffrey Moskot
System Administrator
jef at math.miami.edu