James Carlson wrote:
> Thomas Maier-Komor writes:
>> James Carlson wrote:
>>> Thomas Maier-Komor writes:
>>>> Has anybody got some mdb magic at hand?
>>> Sure. If the PID is 1234 (decimal), do this:
>>>
>>>> 0t1234::pid2proc | ::walk thread | ::findstack
>> Thanks. Looks like a good starter. But what can I do, if
>> $ 0t1234::pid2proc | ::walk thread
>> returns nothing?
>
> Then you have to do:
>
> ::threadlist -v ! less
>
> and slog through the output to find the offender.
>
Great! I found it. It's hanging somewhere in NFS:
PC: cv_wait+0x38 CMD: evolution-2.2
stack pointer for thread 3000247e980: 2a100fd49d1
[ 000002a100fd49d1 cv_wait+0x38() ]
page_lock_es+0x1a8()
pvn_vplist_dirty+0x2a4()
nfs_putpages+0xcc()
nfs3_putpage+0xcc()
nfs3_putpage_commit+0x6c()
nfs3_close+0x12c()
fop_close+0x20()
closef+0x4c()
closeall+0x50()
proc_exit+0x398()
exit+8()
syscall_trap32+0xcc()
Is there a way to resolve this hanging CV? Or can one determine the NFS
resource in question. I guess it might be something caused by one of our
Linux NFS servers...
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
