[Cooker] Problems with nfs 1.2.1

[J.A. Magallón]

Hi...

Since cooker has been reopened I have problems with my systems and a
Solaris 10 NFS server. Systems automount /home/??? via an LDAP mount from
the server.
NFS works fine between cooker boxes (both x86-64), and between Cooker
clients and a Ubuntu server. Also other servers can access the Solaris server.

The only problem is between cooker's and the Sun cluster.
I get this strange error:

Nov  8 21:11:04 lorelei automount[2082]: >> mount.nfs: mount to NFS server 'danae-nfs:/export/home/usuarios/giga/magallon' failed: RPC Error: Unable to send; errno = No buffer space available

lorelei:~# grep danae /proc/mounts
ldap://danae-nfs/automountMapName=auto_home,o=diis /home autofs rw,relatime,fd=7,pgrp=2082,timeout=60,minproto=5,maxproto=5,indirect 0 0

Any ideas ?

BTW, as I wrote this, I downgraded one of the boxes to 2010.0 version
(nfs-utils-1.2.0-2mdv2010.0 + nfs-utils-clients-1.2.0-2mdv2010.0), and
all works fine again !!!
Workaround for now....

--
J.A. Magallon <jamagallon()ono!com>     \               Software is like sex:
                                         \         It's better when it's free
Mandriva Linux release 2010.0 (Cooker) for x86_64
Linux 2.6.29.3-desktop-1mnb (gcc 4.3.2 (GCC) #1 Wed May

[guillomovitch]

J.A. Magallón a écrit :

> Hi...
>
> Since cooker has been reopened I have problems with my systems and a
> Solaris 10 NFS server. Systems automount /home/??? via an LDAP mount from
> the server.
> NFS works fine between cooker boxes (both x86-64), and between Cooker
> clients and a Ubuntu server. Also other servers can access the Solaris server.
>
> The only problem is between cooker's and the Sun cluster.
> I get this strange error:
>
> Nov  8 21:11:04 lorelei automount[2082]: >> mount.nfs: mount to NFS server 'danae-nfs:/export/home/usuarios/giga/magallon' failed: RPC Error: Unable to send; errno = No buffer space available
>
> lorelei:~# grep danae /proc/mounts
> ldap://danae-nfs/automountMapName=auto_home,o=diis /home autofs rw,relatime,fd=7,pgrp=2082,timeout=60,minproto=5,maxproto=5,indirect 0 0
>
> Any ideas ?

 From the changelog:
  * NFS version 4 will be the first version to be tried on client mounts.
    If the server does not support version 4, the mounts will roll
    back to version 3 and then version 2.

Try manual NFS mount, using explicit version option.
--
BOFH excuse #15:

temporary routing anomaly

[J.A. Magallón]

On Mon, 09 Nov 2009 00:14:39 +0100, Guillaume Rousse <Guillaume.Rousse@...> wrote:

> J.A. Magallón a écrit :
> > Hi...
> >
> > Since cooker has been reopened I have problems with my systems and a
> > Solaris 10 NFS server. Systems automount /home/??? via an LDAP mount from
> > the server.
> > NFS works fine between cooker boxes (both x86-64), and between Cooker
> > clients and a Ubuntu server. Also other servers can access the Solaris server.
> >
> > The only problem is between cooker's and the Sun cluster.
> > I get this strange error:
> >
> > Nov  8 21:11:04 lorelei automount[2082]: >> mount.nfs: mount to NFS server 'danae-nfs:/export/home/usuarios/giga/magallon' failed: RPC Error: Unable to send; errno = No buffer space available
> >
> > lorelei:~# grep danae /proc/mounts
> > ldap://danae-nfs/automountMapName=auto_home,o=diis /home autofs rw,relatime,fd=7,pgrp=2082,timeout=60,minproto=5,maxproto=5,indirect 0 0
> >
> > Any ideas ?
>  From the changelog:
>   * NFS version 4 will be the first version to be tried on client mounts.
>     If the server does not support version 4, the mounts will roll
>     back to version 3 and then version 2.
>
> Try manual NFS mount, using explicit version option.

It looks like that is not the problem. mount does not try version 4 if it
is not invoked as mount.nfs4. mount.nfs just tries 3 and 2:

cicely:/mnt# mount.nfs4 danae-nfs:/export/home/extra/grupos/giga /mnt/tst -v
mount.nfs4: timeout set for Mon Nov  9 00:47:21 2009
mount.nfs4: trying text-based options 'addr=xxxxxxx,clientaddr=xxxxxxx'
mount.nfs4: mount(2): Protocol not supported
mount.nfs4: Protocol not supported

But... look:

cicely:/mnt# mount.nfs danae-nfs:/export/home/extra/grupos/giga /mnt/tst -v
mount.nfs: timeout set for Mon Nov  9 00:48:02 2009
mount.nfs: prog 100003, trying vers=3, prot=6
mount.nfs: trying xxx.xxx.xxx.xxx prog 100003 vers 3 prot TCP port 2049
mount.nfs: prog 100005, trying vers=3, prot=17
mount.nfs: portmap query failed: RPC: Unable to send - Resource temporarily unavailable
mount.nfs: mount to NFS server 'danae-nfs:/export/home/extra/grupos/giga' failed: RPC Error: Unable to send; errno = Resource temporarily unavailable

cicely:/mnt# mount.nfs danae-nfs:/export/home/extra/grupos/giga /mnt/tst -v -o tcp
mount.nfs: timeout set for Mon Nov  9 00:48:53 2009
mount.nfs: prog 100003, trying vers=3, prot=6
mount.nfs: trying xxx.xxx.xxx.xxx prog 100003 vers 3 prot TCP port 2049
mount.nfs: prog 100005, trying vers=3, prot=6
mount.nfs: trying xxx.xxx.xxx.xxx prog 100005 vers 3 prot TCP port 48487
mount.nfs: trying text-based options 'addr=xxx.xxx.xxx.xxx,vers=3,proto=tcp,mountvers=3,mountproto=tcp,mountport=48487'
danae-nfs:/export/home/extra/grupos/giga on /mnt/tst type nfs (tcp)

with explicit 'tcp' option it works. It was not needed before...

--
J.A. Magallon <jamagallon()ono!com>     \               Software is like sex:
                                         \         It's better when it's free
Mandriva Linux release 2010.0 (Cooker) for x86_64
Linux 2.6.29.3-desktop-1mnb (gcc 4.3.2 (GCC) #1 Wed May

[Colin Guthrie-8]

'Twas brillig, and J.A. Magallón at 08/11/09 23:56 did gyre and gimble:

> On Mon, 09 Nov 2009 00:14:39 +0100, Guillaume Rousse <Guillaume.Rousse@...> wrote:
>
>> J.A. Magallón a écrit :
>>> Hi...
>>>
>>> Since cooker has been reopened I have problems with my systems and a
>>> Solaris 10 NFS server. Systems automount /home/??? via an LDAP mount from
>>> the server.
>>> NFS works fine between cooker boxes (both x86-64), and between Cooker
>>> clients and a Ubuntu server. Also other servers can access the Solaris server.
>>>
>>> The only problem is between cooker's and the Sun cluster.
>>> I get this strange error:
>>>
>>> Nov  8 21:11:04 lorelei automount[2082]: >> mount.nfs: mount to NFS server 'danae-nfs:/export/home/usuarios/giga/magallon' failed: RPC Error: Unable to send; errno = No buffer space available
>>>
>>> lorelei:~# grep danae /proc/mounts
>>> ldap://danae-nfs/automountMapName=auto_home,o=diis /home autofs rw,relatime,fd=7,pgrp=2082,timeout=60,minproto=5,maxproto=5,indirect 0 0
>>>
>>> Any ideas ?
>>  From the changelog:
>>   * NFS version 4 will be the first version to be tried on client mounts.
>>     If the server does not support version 4, the mounts will roll
>>     back to version 3 and then version 2.
>>
>> Try manual NFS mount, using explicit version option.
>
> It looks like that is not the problem. mount does not try version 4 if it
> is not invoked as mount.nfs4. mount.nfs just tries 3 and 2:
>
> cicely:/mnt# mount.nfs4 danae-nfs:/export/home/extra/grupos/giga /mnt/tst -v
> mount.nfs4: timeout set for Mon Nov  9 00:47:21 2009
> mount.nfs4: trying text-based options 'addr=xxxxxxx,clientaddr=xxxxxxx'
> mount.nfs4: mount(2): Protocol not supported
> mount.nfs4: Protocol not supported
>
> But... look:
>
> cicely:/mnt# mount.nfs danae-nfs:/export/home/extra/grupos/giga /mnt/tst -v
> mount.nfs: timeout set for Mon Nov  9 00:48:02 2009
> mount.nfs: prog 100003, trying vers=3, prot=6
> mount.nfs: trying xxx.xxx.xxx.xxx prog 100003 vers 3 prot TCP port 2049
> mount.nfs: prog 100005, trying vers=3, prot=17
> mount.nfs: portmap query failed: RPC: Unable to send - Resource temporarily unavailable
> mount.nfs: mount to NFS server 'danae-nfs:/export/home/extra/grupos/giga' failed: RPC Error: Unable to send; errno = Resource temporarily unavailable
>
> cicely:/mnt# mount.nfs danae-nfs:/export/home/extra/grupos/giga /mnt/tst -v -o tcp
> mount.nfs: timeout set for Mon Nov  9 00:48:53 2009
> mount.nfs: prog 100003, trying vers=3, prot=6
> mount.nfs: trying xxx.xxx.xxx.xxx prog 100003 vers 3 prot TCP port 2049
> mount.nfs: prog 100005, trying vers=3, prot=6
> mount.nfs: trying xxx.xxx.xxx.xxx prog 100005 vers 3 prot TCP port 48487
> mount.nfs: trying text-based options 'addr=xxx.xxx.xxx.xxx,vers=3,proto=tcp,mountvers=3,mountproto=tcp,mountport=48487'
> danae-nfs:/export/home/extra/grupos/giga on /mnt/tst type nfs (tcp)
>
> with explicit 'tcp' option it works. It was not needed before...
>

Yeah, I'm seeing this and the tcp option fixes it for me too.

Is there a bug report for this yet?

Col

--

Colin Guthrie
cguthrie(at)mandriva.org
http://colin.guthr.ie/

Day Job:
   Tribalogic Limited [http://www.tribalogic.net/]
Open Source:
   Mandriva Linux Contributor [http://www.mandriva.com/]
   PulseAudio Hacker [http://www.pulseaudio.org/]
   Trac Hacker [http://trac.edgewall.org/]

[J.A. Magallón]

On Tue, 17 Nov 2009 17:43:35 +0000, Colin Guthrie <cguthrie@...> wrote:

...
> >
> > It looks like that is not the problem. mount does not try version 4 if it
> > is not invoked as mount.nfs4. mount.nfs just tries 3 and 2:
...

> > But... look:
> >
> > cicely:/mnt# mount.nfs danae-nfs:/export/home/extra/grupos/giga /mnt/tst -v
> > mount.nfs: timeout set for Mon Nov  9 00:48:02 2009
> > mount.nfs: prog 100003, trying vers=3, prot=6
> > mount.nfs: trying xxx.xxx.xxx.xxx prog 100003 vers 3 prot TCP port 2049
> > mount.nfs: prog 100005, trying vers=3, prot=17
> > mount.nfs: portmap query failed: RPC: Unable to send - Resource temporarily unavailable
> > mount.nfs: mount to NFS server 'danae-nfs:/export/home/extra/grupos/giga' failed: RPC Error: Unable to send; errno = Resource temporarily unavailable
> >
> > cicely:/mnt# mount.nfs danae-nfs:/export/home/extra/grupos/giga /mnt/tst -v -o tcp
> > mount.nfs: timeout set for Mon Nov  9 00:48:53 2009
> > mount.nfs: prog 100003, trying vers=3, prot=6
> > mount.nfs: trying xxx.xxx.xxx.xxx prog 100003 vers 3 prot TCP port 2049
> > mount.nfs: prog 100005, trying vers=3, prot=6
> > mount.nfs: trying xxx.xxx.xxx.xxx prog 100005 vers 3 prot TCP port 48487
> > mount.nfs: trying text-based options 'addr=xxx.xxx.xxx.xxx,vers=3,proto=tcp,mountvers=3,mountproto=tcp,mountport=48487'
> > danae-nfs:/export/home/extra/grupos/giga on /mnt/tst type nfs (tcp)
> >
> > with explicit 'tcp' option it works. It was not needed before...
> >
>
> Yeah, I'm seeing this and the tcp option fixes it for me too.
>
> Is there a bug report for this yet?
>

Done.

--
J.A. Magallon <jamagallon()ono!com>     \               Software is like sex:
                                         \         It's better when it's free
Mandriva Linux release 2010.0 (Cooker) for x86_64
Linux 2.6.29.3-desktop-1mnb (gcc 4.3.2 (GCC) #1 Wed May

[Colin Guthrie-8]

'Twas brillig, and J.A. Magallón at 18/11/09 23:17 did gyre and gimble:
> Done.

A bug # would have tasted really good with that message :p
https://qa.mandriva.com/show_bug.cgi?id=55766

Col
--

Colin Guthrie
cguthrie(at)mandriva.org
http://colin.guthr.ie/

Day Job:
   Tribalogic Limited [http://www.tribalogic.net/]
Open Source:
   Mandriva Linux Contributor [http://www.mandriva.com/]
   PulseAudio Hacker [http://www.pulseaudio.org/]
   Trac Hacker [http://trac.edgewall.org/]