Greetings,
For some reason, both Opera and firefox hang regularly when run under
my user account under SXCE b56 (full install). I've had a few other
programs hang too, but they do it quickest.
When I try to attach to the hung process with mdb, here's what happens:
mdb -p 1277
mdb: failed to initialize /lib/libc_db.so.1: libthread_db call failed
unexpectedly
mdb: warning: debugger will only be able to examine raw LWPs
Loading modules: [ ld.so.1 libc.so.1 libuutil.so.1 libavl.so.1 ]
> ::stack
> ::stack
> ::kill
mdb: victim process PID 1277 forcibly terminated
> quit
mdb: failed to dereference symbol: unknown symbol name
>
mdb -p 1277
mdb: cannot debug 1277: unanticipated system error
mdb: failed to initialize target: Error 0
========
Further research seems to indicate that this is related to the zfs
filesystem I am using for my home directory. Why do I believe this?
While logged in as root, not as my user (which has a zfs home directory):
zpool scrub is stuck at "almost" 100%:
pool: tank
state: ONLINE
scrub: scrub in progress, 99.89% done, 0h0m to go
config:
NAME STATE READ WRITE CKSUM
tank ONLINE 0 0 0
c1d0s7 ONLINE 0 0 0
errors: No known data errors
and zfs unmount tank/home hangs forever as well.
mdb shows nothing useful:
bash-3.00# mdb -p 2157
mdb: failed to initialize /lib/libc_db.so.1: libthread_db call failed
unexpectedly
mdb: warning: debugger will only be able to examine raw LWPs
Loading modules: [ ld.so.1 libumem.so.1 libuutil.so.1 libnvpair.so.1
libc.so.1 libavl.so.1 ]
> ::stack
>
====
dtrace gives me nothing useful about the "zfs unmount" process either:
bash-3.00# dtrace -n pid2157:::entry
dtrace: description 'pid2157:::entry' matched 9982 probes
====
If I try to reboot (with "init 6" as root from a console login) once
the problem with one of the programs hanging happens, it hangs forever
and won't restart. I'm guessing a file is still open so the unmount
won't proceed, and there doesn't seem to be any way to force it to do
the unmount.
====
My theory seems to hold true since whenever I run either program as
root instead, I can do so without any problem for hours whereas it
usually only takes 10-15 minutes as my user to have the same issue.
I did have compression set to "on" originally for my zfs filesystems,
but turned it off thinking that might be causing it, and that didn't
help either. I've also tried a "zpool scrub tank" on a fresh boot
which did complete successfully (it never completes if I've had an
application "hang" as I showed earlier).
It isn't exactly reproduceable, but it has been happening every day so far.
My next step was going to be to destroy the pool and start over, but
I'd rather figure out what the problem is before taking such a drastic
measure.
I apologise if this was better sent to another list, but I'm fairly
certain this is zfs related somehow.
Any feedback is welcome, and if I can provide any further information,
I would be most happy to do so.
--
"Less is only more where more is no good." --Frank Lloyd Wright
Shawn Walker, Software and Systems Analyst
binarycrusader at gmail.com - http://binarycrusader.blogspot.com/
