Current-Users archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: Reader / writer lock error: rw_vector_exit: assertion failed: RW_COUNT(rw) != 0
Hi Reinoud,
On Tue, 15 Jul 2008 10:44:48 pm Reinoud Zandijk wrote:
> I kind of replicated it; if i move the dir *inside* the exported fs i get a
> pwd on the _client_ stuck in [tstile] :-S This alone is pretty serious.
Eww ... I'll have to watch that one.
> But if i move the dir *outside* the exported fs i get:
>
> client:
> ------
> luiaard# mkdir 1
> luiaard# cd 1
> luiaard# mkdir 2
> luiaard# cd 2
> luiaard# ls -al
> total 2
> drwxr-xr-x 2 root wheel 512 Jul 15 14:33 .
> drwxr-xr-x 3 root wheel 512 Jul 15 14:33 ..
> luiaard# pwd
> /root/1/2
>
>
> server:
> ------
> heethoofdje# pwd
> /usr/exports/luiaard/root/1
> heethoofdje# mv 2 ~
>
>
> client
> -----
> luiaard# cd ..
> luiaard# ls -al
> <list of ~>
> luiaard# pwd
> Reader / writer lock error: rw_vector_exit: assertion filed : RW_COUNT(rw)
> != 0
> lock address : 0x00000000c7267ab4
> current cpu : 0
> current lwp : 0x00000000c70080a0
> owner/count : 000000000000000000 flags : 000000000000000000
> panic: lock error
>
>
> XXX note that the ls -al on `luiaard' (the client) should list ~/1 but
> lists ~ !!! and thus most likely unlocks/locks the wrong lock and BOEM!
Yep, that's the same thing I'm seeing.
Also, when I stumbled on the above, I don't know how an NFS sever/client is
meant to react when an export dir is moved or deleted but when it's happened,
I've yet to find a way of unmounting/remounting without rebooting the client.
umount won't recognise the stale mount at all.
Sarton
Home |
Main Index |
Thread Index |
Old Index