nfsd panic

[Daniel Braniss]

On Mar 20, 2014, at 9:22 PM, Alexander Motin <mav at freebsd.org> wrote:

> On 20.03.2014 17:08, John Baldwin wrote:
>> On Thursday, March 20, 2014 8:35:37 am Daniel Braniss wrote:
>>> this host has been doing fine, but today it’s constantly crashing in nfsd
>>> it’s exporting a 32TB zfs via nfs, to several hungry hosts
>>> any help is appreciated since this is a production server and some user are
>>> not very happy :-)
>>> 
>>> http://www.cs.huji.ac.il/~danny/core.txt.7
>> 
>> I think the pool->sp_lock mutex is not locked.  Can you go to frame 8 in kgdb
>> and do 'p *m’?

(kgdb) frame 8
#8  0xffffffff808cf25a in _mtx_unlock_sleep (m=0xfffffe002c132400, 
    opts=<value optimized out>, file=<value optimized out>, 
    line=<value optimized out>) at /r+d/stable/9/sys/kern/kern_mutex.c:716
716     /r+d/stable/9/sys/kern/kern_mutex.c: No such file or directory.
        in /r+d/stable/9/sys/kern/kern_mutex.c
(kgdb) p *m
$1 = {lock_object = {lo_name = 0xffffffff80f5d757 "sp_lock", 
    lo_flags = 16973824, lo_data = 0, lo_witness = 0x0}, mtx_lock = 4}
> 
> Daniel, your system looks like updated on February 2, but there was alike bug fixed in stable/9 on February 7 (r261578). Please try to update your system.
> 
I did and now it’s ok again - though will have to wait till traffic picks up again next week
thanks,
	danny

> -- 
> Alexander Motin

thanks to you all,
	danny