How to take down a system to the point of requiring a newfs with one line of C (userland)

[Alexandre "Sunny" Kovalenko]

On Mon, 2008-02-18 at 17:21 +0000, Robert Watson wrote:
> On Mon, 18 Feb 2008, Robert Watson wrote:
> 
> > Hopefully this bug will get resolved shortly, and then we can evaluate if an 
> > errata notice is necessary.
> 
> FYI, I have been unable, thus far, to reproduce it with 150,000 entries in the 
> root of a test file system on an 8.x kernel.  I'm not set up to test 6.x and 
> 7.x currently, and have other obligations tht will prevent me from setting up 
> 6.x and 7.x test images for a few days.
FWIW: I can not reproduce this on the 7.0-RC2:

twinhead# umount /usr/ports/distfiles
twinhead# sync
twinhead# sync
twinhead# sync
twinhead# sync
twinhead# mount /usr/ports/distfiles
twinhead# df -k
Filesystem         1024-blocks      Used     Avail Capacity  Mounted on
/dev/da0s1a            8119416   4714312   2755552    63%    /
devfs                        1         1         0   100%    /dev
/dev/da0s3d            7054514   1032194   5457960    16%    /home
/dev/stripe/shared   103288206  66041510  28983640    69%    /SHARED
/dev/stripe/S0       378425950 116115180 232036694    33%    /STORAGE
procfs                       4         4         0   100%    /proc
/dev/ad4s2            47298314   4314412  39200038
10%    /usr/ports/distfiles
twinhead# cd /usr/ports/distfiles
twinhead# ls | egrep "^[0-9]" | wc -l
   10000
twinhead# ls | wc -l
   10673
twinhead# uname -a
FreeBSD twinhead.rabbitslawn.verizon.net 7.0-RC2 FreeBSD 7.0-RC2 #0: Sat
Feb 16 08:44:12 EST 2008
root at twinhead.rabbitslawn.verizon.net:/usr/obj/usr/src/sys/TWINHEAD
i386

If this makes any difference, this is SMP machine running SMP kernel.

> 
> If people who can reproduce this problem could send kernel stack traces (etc) 
> as a follow-up to the PR, that would be most helpful.  Right now it's sparse 
> on actual debugging data.
> 
> Robert N M Watson
> Computer Laboratory
> University of Cambridge
> _______________________________________________
> freebsd-stable at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-stable
> To unsubscribe, send any mail to "freebsd-stable-unsubscribe at freebsd.org"
-- 
Alexandre "Sunny" Kovalenko (Олександр Коваленко)