I have a server that I use to host ISO images, and mount them so they
are available via network shares. I ran into a problem today, I
temporarily made an ISO image accessible via a md device and mounted it
under /mnt just to check the data on the ISO image. My ISO mount script
ran its updated check while this was there, and hung up because of it.
Now the obvious solution is to fix my script, but as I am baffled as to
why it hung-up, it does do a query on md devices, and will try to
dismount any that are mounted if they don't match its criteria, and
delete the md device as well. However as it would have found this one
in use, it should have just returned a failure and continued on.
However it didn't and then when I tried to manually umount it, my umount
command hung as well. What I was left with was two umount commands
attempting to umount /dev/md1000 both stuck, they wouldn't respond to a
kill -9 I couldn't use mdconfig -d -u 1000 to delete the md device, even
with a -o force (had yet another process stuck). After 1 hour all
processes were still hung, killing the shell left them zombied, but
still hung there, tying up the md device.
I was left with rebooting the server, until I can figure out why my
script broke and didn't just error and continue the remaining checks in
its list. Does anyone have any idea how if this happens again to kill
these hung-up umount processes, without rebooting the server?
This could be a fun one to fix, because so far attempting to duplicate
the problem hasn't worked, think I just have to time something just
right, or wrong depended how you look at it.
--
Thanks,
Dean E. Weimer
http://www.dweimer.net/
_______________________________________________
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
