NFS errors

NFS errors

[Carsten Otto]

[-- Attachment #1: Type: text/plain, Size: 3103 bytes --]

Hello!

My computer uses another server to boot and run with NFS root.
During normal work I notice strange NFS errors.
Please tell me what causes these errors and how I can resolve them.

Example 1:
I executed these commands directly after another.

cotto@carsten-otto:/home/cotto/.mozilla/firefox$ l
ls: .: Permission denied
cotto@carsten-otto:/home/cotto/.mozilla/firefox$ l
total 8.0K
drwx------ 6 cotto users 4.0K Jul 24 21:03 eir51v0r.default
?--------- ? ?     ?        ?            ? la2v5nug.default
?--------- ? ?     ?        ?            ? pluginreg.dat
-rw-r--r-- 1 cotto users   94 Jul 24 21:03 profiles.ini
?--------- ? ?     ?        ?            ? v1q37ws3.default
cotto@carsten-otto:/home/cotto/.mozilla/firefox$ l
total 20K
drwx------ 6 cotto users 4.0K Jul 24 21:03 eir51v0r.default
drwx------ 3 cotto users 4.0K Jul 24 21:03 la2v5nug.default
-rw------- 1 cotto users  286 Jul 24 20:57 pluginreg.dat
-rw-r--r-- 1 cotto users   94 Jul 24 21:03 profiles.ini
drwx------ 6 cotto users 4.0K Jul 24 21:04 v1q37ws3.default

Example 2:
GAIM shows "Could not read file: somewhere.txt" when scrolling through
the list of archived conversations.

Example 3:
Firefox crashes a lot during normal surfing, sometimes even the profile
gets lost.

Example 4:
I get a lot of "do_vfs_lock: VFS is out of sync with lock manager!"
during the day.

Example 5:
Some ./configure && make runs do not complete, because files are
missing. Running the command in a chroot locally on the server (which
has the disk drives) everything works.

The client (without disk drives) uses 2.6.17.6, but I have this problem
with several kernel versions.
The kernel is running 2.6.18-rc1-mm2, but here again the problem occurs
with several versions.
The client runs Debian testing, the server runs Debian stable (Sarge).
The network connection in between is:
(Client) Intel e1000 PCI <-> HP Procurve 2824 (Gigabit) <-> Intel e1000 PCIe (Server)

x:~# rpcinfo -p server
   Program Vers Proto   Port
    100000    2   tcp    111  portmapper
    100000    2   udp    111  portmapper
    100003    2   udp   2049  nfs
    100003    3   udp   2049  nfs
    100003    4   udp   2049  nfs
    100003    2   tcp   2049  nfs
    100003    3   tcp   2049  nfs
    100003    4   tcp   2049  nfs
    100021    1   udp  32768  nlockmgr
    100021    3   udp  32768  nlockmgr
    100021    4   udp  32768  nlockmgr
    100021    1   tcp  43642  nlockmgr
    100021    3   tcp  43642  nlockmgr
    100021    4   tcp  43642  nlockmgr
    100005    1   udp    779  mountd
    100005    1   tcp    782  mountd
    100005    2   udp    779  mountd
    100005    2   tcp    782  mountd
    100005    3   udp    779  mountd
    100005    3   tcp    782  mountd
    100024    1   udp    877  status
    100024    1   tcp    880  status

I enabled nfs4 in the kernels, but do not use it yet
(as a lot of changes are needed in userspace).

mount shows:
server:/export on / type nfs (rw,soft,tcp,intr)

Thanks a lot,
-- 
Carsten Otto
c-otto@gmx.de
www.c-otto.de

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: NFS errors

[Trond Myklebust]

On Mon, 2006-07-24 at 21:15 +0200, Carsten Otto wrote:
> Hello!
> 
> My computer uses another server to boot and run with NFS root.
> During normal work I notice strange NFS errors.
> Please tell me what causes these errors and how I can resolve them.
> 
> Example 1:
> I executed these commands directly after another.
> 
> cotto@carsten-otto:/home/cotto/.mozilla/firefox$ l
> ls: .: Permission denied
> cotto@carsten-otto:/home/cotto/.mozilla/firefox$ l
> total 8.0K
> drwx------ 6 cotto users 4.0K Jul 24 21:03 eir51v0r.default
> ?--------- ? ?     ?        ?            ? la2v5nug.default
> ?--------- ? ?     ?        ?            ? pluginreg.dat
> -rw-r--r-- 1 cotto users   94 Jul 24 21:03 profiles.ini
> ?--------- ? ?     ?        ?            ? v1q37ws3.default
> cotto@carsten-otto:/home/cotto/.mozilla/firefox$ l
> total 20K
> drwx------ 6 cotto users 4.0K Jul 24 21:03 eir51v0r.default
> drwx------ 3 cotto users 4.0K Jul 24 21:03 la2v5nug.default
> -rw------- 1 cotto users  286 Jul 24 20:57 pluginreg.dat
> -rw-r--r-- 1 cotto users   94 Jul 24 21:03 profiles.ini
> drwx------ 6 cotto users 4.0K Jul 24 21:04 v1q37ws3.default
> 
> Example 2:
> GAIM shows "Could not read file: somewhere.txt" when scrolling through
> the list of archived conversations.
> 
> Example 3:
> Firefox crashes a lot during normal surfing, sometimes even the profile
> gets lost.
> 
> Example 4:
> I get a lot of "do_vfs_lock: VFS is out of sync with lock manager!"
> during the day.
> 
> Example 5:
> Some ./configure && make runs do not complete, because files are
> missing. Running the command in a chroot locally on the server (which
> has the disk drives) everything works.
> 
> The client (without disk drives) uses 2.6.17.6, but I have this problem
> with several kernel versions.
> The kernel is running 2.6.18-rc1-mm2, but here again the problem occurs
> with several versions.
> The client runs Debian testing, the server runs Debian stable (Sarge).
> The network connection in between is:
> (Client) Intel e1000 PCI <-> HP Procurve 2824 (Gigabit) <-> Intel e1000 PCIe (Server)
> 
> x:~# rpcinfo -p server
>    Program Vers Proto   Port
>     100000    2   tcp    111  portmapper
>     100000    2   udp    111  portmapper
>     100003    2   udp   2049  nfs
>     100003    3   udp   2049  nfs
>     100003    4   udp   2049  nfs
>     100003    2   tcp   2049  nfs
>     100003    3   tcp   2049  nfs
>     100003    4   tcp   2049  nfs
>     100021    1   udp  32768  nlockmgr
>     100021    3   udp  32768  nlockmgr
>     100021    4   udp  32768  nlockmgr
>     100021    1   tcp  43642  nlockmgr
>     100021    3   tcp  43642  nlockmgr
>     100021    4   tcp  43642  nlockmgr
>     100005    1   udp    779  mountd
>     100005    1   tcp    782  mountd
>     100005    2   udp    779  mountd
>     100005    2   tcp    782  mountd
>     100005    3   udp    779  mountd
>     100005    3   tcp    782  mountd
>     100024    1   udp    877  status
>     100024    1   tcp    880  status
> 
> I enabled nfs4 in the kernels, but do not use it yet
> (as a lot of changes are needed in userspace).
> 
> mount shows:
> server:/export on / type nfs (rw,soft,tcp,intr)
> 
> Thanks a lot,

Have you set 'no_subtree_check' in your export options on the server? If
not, please try doing so.

cheers,
  Trond

Re: NFS errors

[Janos Farkas]

Hi!

On 2006-07-24 at 15:46:48, Trond Myklebust wrote:
> Have you set 'no_subtree_check' in your export options on the server? If
> not, please try doing so.

Oh.  But Neil almost told me that :)  That seems to fix a similar
problem for me at least:

http://thread.gmane.org/gmane.linux.kernel/426931

Now I'm not seeing any weirdness between two 2.6.18-rc2 hosts.  (No
spurious "Reducing readahead size to 28/4/0K" either).  I still don't
get how that patch that Neil mentioned did not affect 2.6.17 (which had
that), but does effect any 2.6.18-rcX.

Janos

Re: NFS errors

[Carsten Otto]

[-- Attachment #1: Type: text/plain, Size: 288 bytes --]

On Mon, Jul 24, 2006 at 03:46:48PM -0400, Trond Myklebust wrote:
> Have you set 'no_subtree_check' in your export options on the server? If
> not, please try doing so.

No problem so far (in the last few days), seems to work. Thanks!
-- 
Carsten Otto
c-otto@gmx.de
www.c-otto.de

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]