8:51 a.m.
Hi -
Just upgraded from 2.6.21-1.3228.fc7, and now seeing problems with NFS. I originally
believed the problem was with the automounter, but it looks like it might be NFS instead.
Rebooting with the old kernel solves the problem.
In this example, I'll create a new mountpoint, and try to mount my home directory. The
directory is being served from a SUN Cluster 2.1 server.
# rm -rf /h
# mkdir -p /h/jminer
# mount superfly:/superfly/vol02/home/jminer /h/jminer
mount.nfs: /h/jminer is already mounted or busy
# echo $?
32
Now, if I try to mount the directory from the physical node, the mount suceeds.
# mount fly_a:/superfly/vol02/home/jminer /h/jminer
# echo $?
0
I saw similar problems with an earlier FC release, but it was related to either the autofs
or nfs-utils package.
9:11 a.m.
-----Original Message-----
From: fedora-list-bounces(a)redhat.com on behalf of Andy Green
Sent: Thu 07/26/2007 09:54 AM
To: For users of Fedora
Cc:
Subject: Re: Kernel 2.6.22.1-27.fc7 breaks NFS
Somebody in the thread at some point said:
I saw similar problems with an earlier FC release, but it was
related
to either the autofs or nfs-utils package.
Just to say automount is broken currently here on that kernel... takes
up 30% of CPU doing nothing.
---------------------------
That is not happening here. My automount daemon is idle.
9:52 a.m.
-----Original Message-----
From: fedora-list-bounces(a)redhat.com on behalf of Miner, Jonathan W (CSC) (US SSA)
Sent: Thu 07/26/2007 09:51 AM
To: fedora-list(a)redhat.com
Cc:
Subject: Kernel 2.6.22.1-27.fc7 breaks NFS
Hi -
Just upgraded from 2.6.21-1.3228.fc7, and now seeing problems with NFS. I originally
believed the problem was with the automounter, but it looks like it might be NFS instead.
Rebooting with the old kernel solves the problem.
In this example, I'll create a new mountpoint, and try to mount my home directory. The
directory is being served from a SUN Cluster 2.1 server.
# rm -rf /h
# mkdir -p /h/jminer
# mount superfly:/superfly/vol02/home/jminer /h/jminer
mount.nfs: /h/jminer is already mounted or busy
# echo $?
32
Now, if I try to mount the directory from the physical node, the mount suceeds.
# mount fly_a:/superfly/vol02/home/jminer /h/jminer
# echo $?
0
I saw similar problems with an earlier FC release, but it was related to either the autofs
or nfs-utils package.
--------------------------------------
I just learned something else. If I stop the autofs (/etc/init.d/autofs stop), then the
original mount command works fine. If autofs is restarted, the mount command fails again.
10 a.m.
Somebody in the thread at some point said:
I just learned something else. If I stop the autofs
(/etc/init.d/autofs stop), then the original mount command works
fine. If autofs is restarted, the mount command fails again.
autofs service == automount process
# yum --enablerepo development update autofs
I updated mine to the one from the development repo, on restart it
didn't take up 30% of CPU... maybe worth a try...
-Andy
10:44 a.m.
-----Original Message-----
From: fedora-list-bounces(a)redhat.com on behalf of Andy Green
Sent: Thu 07/26/2007 11:00 AM
To: For users of Fedora
Cc:
Subject: Re: Kernel 2.6.22.1-27.fc7 breaks NFS
Somebody in the thread at some point said:
I just learned something else. If I stop the autofs
(/etc/init.d/autofs stop), then the original mount command works
fine. If autofs is restarted, the mount command fails again.
autofs service == automount process
# yum --enablerepo development update autofs
I updated mine to the one from the development repo, on restart it
didn't take up 30% of CPU... maybe worth a try...
-----------------------------------------------------------------------
Updating autofs did not fix the problem.
4:06 p.m.
Miner, Jonathan W (CSC) (US SSA) wrote:
-----Original Message-----
From: fedora-list-bounces(a)redhat.com on behalf of Miner, Jonathan W (CSC) (US SSA)
Sent: Thu 07/26/2007 09:51 AM
To: fedora-list(a)redhat.com
Cc:
Subject: Kernel 2.6.22.1-27.fc7 breaks NFS
Hi -
Just upgraded from 2.6.21-1.3228.fc7, and now seeing problems with NFS. I originally
believed the problem was with the automounter, but it looks like it might be NFS instead.
Rebooting with the old kernel solves the problem.
In this example, I'll create a new mountpoint, and try to mount my home directory.
The directory is being served from a SUN Cluster 2.1 server.
# rm -rf /h
# mkdir -p /h/jminer
# mount superfly:/superfly/vol02/home/jminer /h/jminer
mount.nfs: /h/jminer is already mounted or busy
# echo $?
32
Now, if I try to mount the directory from the physical node, the mount suceeds.
# mount fly_a:/superfly/vol02/home/jminer /h/jminer
# echo $?
0
I saw similar problems with an earlier FC release, but it was related to either the
autofs or nfs-utils package.
--------------------------------------
I just learned something else. If I stop the autofs (/etc/init.d/autofs stop), then the
original mount command works fine. If autofs is restarted, the mount command fails again.
If you don't need automount don't run it. If you do, please post your
config file. I have seen similar when a lower directory is mounted, or
in the automount config, like mounting superfly:/superfly/vol02/home as
an example.
Some portion of problems with new kernels result from added error
checking in the new kernel telling you that you have been doing it wrong
and getting away with it. I'm not saying that happened here, but this
sounds a lot as if automount has some part of that mount in config, or
being used, or otherwise referenced.
You might try running automount, check that the mount fails, then run
lsof and see if automount is using something related.
--
Bill Davidsen <davidsen(a)tmr.com>
"We have more to fear from the bungling of the incompetent than from
the machinations of the wicked." - from Slashdot
3:15 a.m.
We're seeing the same thing here. If I use a 2.6.20 kernel for instance, I have no
problems. If I use the latest released FC6 kernel (2.6.22.4-45) it gets the "is
already mounted or busy".
To note, this isn't an issue with "automount" (autofs), but rather an issue
with NFS itself. You can't manually mount either (the original example is vanilla NFS
too incidentally). Automount is just an abstraction on top (which yes, is not surprisingly
broken too).
--
This is an email sent via The Fedora Community Portal https://fcp.surfsite.org
https://fcp.surfsite.org/modules/newbb/viewtopic.php?post_id=201008&t...
If you think, this is spam, please report this to webmaster(a)fcp.surfsite.org and/or blame
Matt.Fahrner(a)coat.com.
5:57 a.m.
On Tue, 11 Sep 2007, [UTF-8] scathew wrote:
We're seeing the same thing here. If I use a 2.6.20 kernel for
instance,
I have no problems. If I use the latest released FC6 kernel
(2.6.22.4-45) it gets the "is already mounted or busy".
To note, this isn't an issue with "automount" (autofs), but rather an
issue with NFS itself. You can't manually mount either (the original
example is vanilla NFS too incidentally). Automount is just an
abstraction on top (which yes, is not surprisingly broken too).
This is bug 250597
https://bugzilla.redhat.com/show_bug.cgi?id=250597
backported to FC6. The fix is to update nfs-utils, then add the
nosharecache option to the nfs mount configuration, eg. by adding
-Onosharecache to the OPTIONS line in /etc/sysconfig/autofs .
Michael Young
1:47 a.m.
Excellent! Totally did the trick!
Thank you VERY MUCH!!!
--
This is an email sent via The Fedora Community Portal https://fcp.surfsite.org
https://fcp.surfsite.org/modules/newbb/viewtopic.php?post_id=201168&t...
If you think, this is spam, please report this to webmaster(a)fcp.surfsite.org and/or blame
Matt.Fahrner(a)coat.com.
3:19 p.m.
I was having same problems with 2.6.22.7-57.fc6 kernel. I added nosharecache
option and was able to
mount the sub-directories, but now when I mount the sub-directories it
crashes rpcbind on my
NFS server. We have Auspex NS3000 NFS server.
I have nfs-utils-1.0.10-14.fc6 and nfs-utils-lib-1.0.8-7.2 installed on the
system.
When rpcbind crashes the NFS access on Unix (AIX, HP-UX, various flavors of
Linux) slows down. I also share the NFS filesystem with PCs using samba and
the PCs cannot access the files.
M A Young wrote:
On Tue, 11 Sep 2007, [UTF-8] scathew wrote:
> We're seeing the same thing here. If I use a 2.6.20 kernel for instance,
> I have no problems. If I use the latest released FC6 kernel
> (2.6.22.4-45) it gets the "is already mounted or busy".
>
> To note, this isn't an issue with "automount" (autofs), but rather an
> issue with NFS itself. You can't manually mount either (the original
> example is vanilla NFS too incidentally). Automount is just an
> abstraction on top (which yes, is not surprisingly broken too).
This is bug 250597
https://bugzilla.redhat.com/show_bug.cgi?id=250597
backported to FC6. The fix is to update nfs-utils, then add the
nosharecache option to the nfs mount configuration, eg. by adding
-Onosharecache to the OPTIONS line in /etc/sysconfig/autofs .
Michael Young
--
fedora-list mailing list
fedora-list(a)redhat.com
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
--
View this message in context:
http://www.nabble.com/Kernel-2.6.22.1-27.fc7-breaks-NFS-tf4151774.html#a1...
Sent from the Fedora List mailing list archive at Nabble.com.
6029
days inactive
6112
days old
10 comments
6 participants
participants (6)
-
Andy Green -
Bill Davidsen -
M A Young -
Miner, Jonathan W (US SSA) -
potat0 -
shahhe