db->put: Invalid argument?

[Ron Jarrell]

At 09:18 AM 7/4/01 +0100, Alex Kiernan wrote:

>Ron Jarrell <jarrell at solaris.cc.vt.edu> writes:
>
>> At 08:34 AM 7/3/01 +0100, Alex Kiernan wrote:
>> >Ron Jarrell <jarrell at solaris.cc.vt.edu> writes:
>> >
>> >> At 09:34 AM 6/28/01 +0100, Alex Kiernan wrote:
>> >> 
>> >> >Ron Jarrell <jarrell at solaris.cc.vt.edu> writes:
>> >> >
>> >> >> At 09:38 AM 6/27/01 +0100, Alex Kiernan wrote:
>> >> >> >jarrell at solaris.cc.vt.edu writes:
>> >> >> >
>> >> >> >> Ok, I just had my 2.3.2 system, which has been running fine for about
>> >> >> >> a week, shudder to a halt.  innd is wedged up tight.  It appears to
>> >> >> >> have died right about the time this appeared in news.err:
>> >> >> >> 
>> >> >> >
>> >> >> >truss it & see where it gets stuck (but I'm so behind on reading stuff
>> >> >> >you've probably already killed it).
>> >> >> 
>> >> >> Yea, I did.  However, I did actually truss it.  It was stuck in a lightweight
>> >> >> process mutex lock.
>> >> >
>> >> >Forgot to ask - what version of BDB are you running with?
>> >> 
>> >> 3.2
>> >> 
>> >
>> >3.2.9?
>> 
>> Yea, sorry, 3.2.9.
>
>Just a shot in the dark :-(
>
>Well patched kernel too? I'm trying to figure why I'm not seeing the
>problem anymore (and I was in the past with some regularity).

Yea.  Right before we brought up the server, we updated the recommended
solaris 8 patch set.  There's some of the optionals we didn't install, but nothing
that seemed extremely pertinent.

Did you used to have problems with getting the low article count be -1 in
some groups?  I had a couple do that to me early on (and I've seen others
have the problem)  Maybe the overview db had a bad record in it because
of that, which has since rolled off because the cnfs buffers cycled, so I'm
not seeing the error any more?