Rick Kelly wrote:
> sigsegv@rambler.ru said:
> 
> 
>>Well no, it happened on i386 once, but the machine had only 32MB of RAM, 
>>and used swap quite often, so I migrated BIND to sparc64 with 512MB of 
>>RAM, and a day later I found out 'named' daemon somehow terminated 
>>again. I've restarted it, and it's been running for about a 48 hours... 
>>still waiting to see how long it will last.
> 
> 
> So this seems to point to threaded app gets swapped, then dies, or machine
> dies.

Yeah, this is the famous bug that many people are aware about, in fact, 
when I was running netbsd-2.0 on sparc64 I used to get kernel panics 
attributed to this bug all the time. I'm running netbsd-current and it 
seems much more stable, well at least the kernel does not crash.

May be this issue with BIND has something to do with that bug, or its 
relics, I don't know.

Well it happened again...

Mar 16 23:31:22 u10 /netbsd: pid 25023 (named), uid 14: exited on signal 
10 (core not dumped, err = 1)

All other applications dump core, but there is no core dump under 
/var/chroot/named is there a way to make it dump core next time?