Joe Holden wrote:
Ma wrote:
I have almost the same reboots on my server. :( And it may reboot
serval times a day. I'd like to know how to get crash dumps? What is
added in your rc.conf?
--
Ma Jie
Hi, I used the rc.conf values as set out in rc,conf(5)
[EMAIL PROTECTED] grep crash /etc/defaults/rc.conf
dumpdev="NO" # Device to crashdump to (device name, AUTO, or
NO).
dumpdir="/var/crash" # Directory where crash dumps are to be stored
I set dumpdev to my swap partition, although nothing got dumped,
although I did upgrade it to -CURRENT as i'd exhausted everything else,
and its not crashed since, not wishing to flame/troll/whatever, but 6.x
seems to have a fair few issues.
I posted a dump of a crash I had with a recent RELENG_6 computer a few
weeks ago. It was one of those fairly useless dumps where the program
counter seems to be completely meaningless.
I also had similar behavior about a year ago, which seemed to be
produced by a bug that is now fixed. (That one I could make it happen
at will by running a multithreaded program on a 4 processor system, and
"top -s0" at the same time.) I do recall the nature of the crash is
that sometimes it produced dumps, and sometimes it didn't. And when it
did they were useless.
The crash seemed always to be with a page fault. I experienced it on
two computers, both with SMP.
Anyway, what I am saying is that all these people might be seeing a
genuine problem that by its very nature does not produce meaningful dumps.
If this problem is real, perhaps one solution is to put out release 6.2
fairly soon, then lots of people will try it and experience this
problem. Then you guys will have lots of evidence to work with. Then
when you fix it, put out release 6.3 much sooner than otherwise
anticipated. (I should add that my wife accuses me of having humor that
is too dry.)
Stephen
_______________________________________________
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "[EMAIL PROTECTED]"
