dump is stuck

[Poul-Henning Kamp]

In message <E1CuXry-000JJd-OV at cs1.cs.huji.ac.il>, Danny Braniss writes:
>> In message <20050128144733.GA91982 at green.homeunix.org>, Brian Fundakowski Feldman writes:
>> >On Fri, Jan 28, 2005 at 09:08:54AM +0200, Danny Braniss wrote:
>> >> hi,
>> >> 	while running 'dump 0f - /dist | restore rf -'
>> >> the dump proc. got stuck, it seems it's waiting on some lock:
>> >> 
>> >> UID   PID  PPID CPU PRI NI   VSZ  RSS MWCHAN STAT  TT       TIME COMMAND
>> >> 
>> >>  0 30924 30922   0   4  0  3396 2852 sbwait T     p1    1:00.88 dump: 
>> >> /dev/amrd0s3h: ...
>> >>  0 30925 30924   1  -8  0  3268 2784 physrd TL    p1    0:53.84 dump 0f - 
>> >> /dist (dump)
>> >>  0 30926 30924   1  20  0  3268 2784 pause  T     p1    0:53.69 dump 0f - 
>> >> /dist (dump)
>> >>  0 30927 30924   1  20  0  3268 2784 pause  T     p1    0:54.12 dump 0f - 
>> >> /dist (dump)
>> >> 
>> >> (this is  5.3-STABLE, cvs'ed about a week ago, and it's a SMP system).
>> >> how can i find which lock? or who is holding it?
>> >
>> >Is the one in physrd not actually reading anything from the disk right
>> >now?  I would suspect that should be how you really determine if it's
>> >hung or not.  You should be able to see how long it's been waiting
>> >and how long it's due to wait still, using kgdb.
>> 
>> Check also with gstat(8) if there is I/O activity going on and/or if any
>> I/O requests are stuck.
>
>it's stuck. i.e. not doing anything. i've been monitoring it via
>iostat, and nothing is moving, nada, the machine is very idle :-(

Please use gstat(8) and look for stuck I/O requests.


-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.