[OpenIndiana-discuss] samba keeps dumping core on oi_147

[Christopher Chan]

On Wednesday, October 27, 2010 11:37 AM, Mike Gerdts wrote:
> On Tue, Oct 26, 2010 at 10:25 PM, Christopher Chan
> <christopher.chan at bradbury.edu.hk>  wrote:
>> Ah, so that's where I can tell SMF to leave things alone. I concur. Running
>> smbd unmanaged has resulted in zero samba downtime. As for the core dumps...
>>
>> In the logs: smbd[22213] setid process, core not dumped:
>>
>> coreadm
>>      global core file pattern: /var/cores/%d/%f.%p.%n
>
> This directory needs to exist.  That is, /var/cores/smbd needs to
> exist with this pattern.  If the directory doesn't exist, you should
> be getting messages in /var/adm/messages (depending on syslog setup)
> that says that the core dump failed because the directory doesn't
> exist.  I tend to avoid using / in the pattern after the first %.
> That is, I'd be more inclined to use /var/cores/%d.%f.%p.%n.  If you
> are aware of what you are doing, the pattern than you use can be
> effective for selectively getting core dumps from only certain
> processes.
>
> I commonly use this technique (core file pattern puts core files in
> /var/cores, but /var/cores doesn't usually exist) to get a log entry
> in /var/adm/messages saying that something tried to dump core without
> filling /var.  This has helped me many times when application admins
> come to me to tell me that somebody is killing their processes.  If I
> care to capture some core files, I'll create the /var/cores directory
> until I get however many I am looking for.

Hmm...a closer look at the log tells me that there have been no core 
dumps since yesterday after I disabled the samba service and ran smbd 
manually. Last entry was at Oct 26 11:30.

But I still have Unclean shutdown entries in log.smbd and associated brl 
and lock database clean ups...