All of lore.kernel.org
 help / color / mirror / Atom feed
* [3.6.11] NFS: Cache request denied due to non-unique superblock keys.
@ 2012-12-19 18:58 Paweł Sikora
  2012-12-19 19:30 ` Jim Rees
  2012-12-19 19:35 ` Myklebust, Trond
  0 siblings, 2 replies; 6+ messages in thread
From: Paweł Sikora @ 2012-12-19 18:58 UTC (permalink / raw)
  To: linux-nfs

Hi,

i'm testing the 3.6.11 kernel on some small internal nfs network (1 server <-> 2 clients)
and observing such info in dmesg on the client side. could you put some light on this dmesg entry?
is it a bug or misconfiguration? i can provide more information about nfs configuration
if necessary but denying caching doesn't look nicely.

i'm using following nfs related packages:

autofs-5.0.5-1.x86_64
libnfsidmap-0.25-4.x86_64
nfs-utils-1.2.7-1.x86_64
nfs-utils-clients-1.2.7-1.x86_64
nfs-utils-common-1.2.7-1.x86_64

BR,
Paweł.

please CC me on reply.


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [3.6.11] NFS: Cache request denied due to non-unique superblock keys.
  2012-12-19 18:58 [3.6.11] NFS: Cache request denied due to non-unique superblock keys Paweł Sikora
@ 2012-12-19 19:30 ` Jim Rees
  2012-12-19 19:35 ` Myklebust, Trond
  1 sibling, 0 replies; 6+ messages in thread
From: Jim Rees @ 2012-12-19 19:30 UTC (permalink / raw)
  To: Paweł Sikora; +Cc: linux-nfs

Paweł Sikora wrote:

  Hi,
  
  i'm testing the 3.6.11 kernel on some small internal nfs network (1 server <-> 2 clients)
  and observing such info in dmesg on the client side. could you put some
  light on this dmesg entry?

Are you mounting the same file system more than once? Why? I thought I
remembered a mount option to uniquify the mounts but I can't find it now.

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [3.6.11] NFS: Cache request denied due to non-unique superblock keys.
  2012-12-19 18:58 [3.6.11] NFS: Cache request denied due to non-unique superblock keys Paweł Sikora
  2012-12-19 19:30 ` Jim Rees
@ 2012-12-19 19:35 ` Myklebust, Trond
  2012-12-19 21:16   ` Paweł Sikora
  1 sibling, 1 reply; 6+ messages in thread
From: Myklebust, Trond @ 2012-12-19 19:35 UTC (permalink / raw)
  To: Paweł Sikora; +Cc: linux-nfs@vger.kernel.org

On Wed, 2012-12-19 at 19:58 +0100, Paweł Sikora wrote:
> Hi,
> 
> i'm testing the 3.6.11 kernel on some small internal nfs network (1 server <-> 2 clients)
> and observing such info in dmesg on the client side. could you put some light on this dmesg entry?
> is it a bug or misconfiguration? i can provide more information about nfs configuration
> if necessary but denying caching doesn't look nicely.
> 
> i'm using following nfs related packages:
> 
> autofs-5.0.5-1.x86_64
> libnfsidmap-0.25-4.x86_64
> nfs-utils-1.2.7-1.x86_64
> nfs-utils-clients-1.2.7-1.x86_64
> nfs-utils-common-1.2.7-1.x86_64
> 
> BR,
> Paweł.

It's hard to guess when you don't tell us anything about what you are
doing to trigger this message, but it sounds as if you are trying to use
the cachefs mount options incorrectly.

-- 
Trond Myklebust
Linux NFS client maintainer

NetApp
Trond.Myklebust@netapp.com
www.netapp.com

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [3.6.11] NFS: Cache request denied due to non-unique superblock keys.
  2012-12-19 19:35 ` Myklebust, Trond
@ 2012-12-19 21:16   ` Paweł Sikora
  2012-12-19 22:06     ` Myklebust, Trond
  0 siblings, 1 reply; 6+ messages in thread
From: Paweł Sikora @ 2012-12-19 21:16 UTC (permalink / raw)
  To: Myklebust, Trond; +Cc: linux-nfs@vger.kernel.org

On Wednesday 19 of December 2012 19:35:50 Myklebust, Trond wrote:
> On Wed, 2012-12-19 at 19:58 +-0100, Pawe+AUI- Sikora wrote:
> +AD4- Hi,
> +AD4- 
> +AD4- i'm testing the 3.6.11 kernel on some small internal nfs network (1 server +ADw--+AD4- 2 clients)
> +AD4- and observing such info in dmesg on the client side. could you put some light on this dmesg entry?
> +AD4- is it a bug or misconfiguration? i can provide more information about nfs configuration
> +AD4- if necessary but denying caching doesn't look nicely.
> +AD4- 
> +AD4- i'm using following nfs related packages:
> +AD4- 
> +AD4- autofs-5.0.5-1.x86+AF8-64
> +AD4- libnfsidmap-0.25-4.x86+AF8-64
> +AD4- nfs-utils-1.2.7-1.x86+AF8-64
> +AD4- nfs-utils-clients-1.2.7-1.x86+AF8-64
> +AD4- nfs-utils-common-1.2.7-1.x86+AF8-64
> +AD4- 
> +AD4- BR,
> +AD4- Pawe+AUI-.
> 
> It's hard to guess when you don't tell us anything about what you are
> doing to trigger this message, but it sounds as if you are trying to use
> the cachefs mount options incorrectly.

on the server side there're following exports:

# exportfs -v
/home/atest/farm-local
                <world>(rw,async,wdelay,nohide,crossmnt,no_root_squash,no_subtree_check,insecure_locks)
/home/atest/farm
                <world>(rw,async,wdelay,nohide,crossmnt,no_root_squash,no_subtree_check,insecure_locks)
/home/users     <world>(rw,no_root_squash,no_subtree_check,insecure_locks)

in the exported /home/atest/* there're crossmounted/binded directories:

# fstab
/home/atest/farm-local/results  /home/atest/farm/results        none bind
/home/services/ftp              /var/lib/ftp                    none bind
/home/services/ftp              /home/atest/farm-local/ftp      none bind

# df -h
Filesystem                   Size  Used Avail Use% Mounted on
rootfs                       7.6G  4.5G  2.7G  63% /
/dev/md0                     7.6G  4.5G  2.7G  63% /
run                           32G  264K   32G   1% /run
devtmpfs                      32G     0   32G   0% /dev
/dev/md1                     7.2T  794G  6.4T  11% /home
none                          32G     0   32G   0% /dev/shm
/dev/md1                     7.2T  794G  6.4T  11% /home/atest/farm/results
/dev/md1                     7.2T  794G  6.4T  11% /var/lib/ftp
/dev/md1                     7.2T  794G  6.4T  11% /home/atest/farm-local/ftp


on the client side there's only a simple autofs configuration:

# cat /etc/autofs/auto.server
farm           -rw,soft,rsize=32768,wsize=32768,intr,vers=4,async,bg,noatime,retry=30,retrans=10        192.168.2.100:/home/atest/farm-local
home           -rw,soft,vers=4        192.168.2.100:/home/users


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [3.6.11] NFS: Cache request denied due to non-unique superblock keys.
  2012-12-19 21:16   ` Paweł Sikora
@ 2012-12-19 22:06     ` Myklebust, Trond
  2012-12-20 13:42       ` David Howells
  0 siblings, 1 reply; 6+ messages in thread
From: Myklebust, Trond @ 2012-12-19 22:06 UTC (permalink / raw)
  To: Paweł Sikora; +Cc: linux-nfs@vger.kernel.org, David Howells

On Wed, 2012-12-19 at 22:16 +0100, Paweł Sikora wrote:
> On Wednesday 19 of December 2012 19:35:50 Myklebust, Trond wrote:
> > On Wed, 2012-12-19 at 19:58 +-0100, Pawe+AUI- Sikora wrote:
> > +AD4- Hi,
> > +AD4- 
> > +AD4- i'm testing the 3.6.11 kernel on some small internal nfs network (1 server +ADw--+AD4- 2 clients)
> > +AD4- and observing such info in dmesg on the client side. could you put some light on this dmesg entry?
> > +AD4- is it a bug or misconfiguration? i can provide more information about nfs configuration
> > +AD4- if necessary but denying caching doesn't look nicely.
> > +AD4- 
> > +AD4- i'm using following nfs related packages:
> > +AD4- 
> > +AD4- autofs-5.0.5-1.x86+AF8-64
> > +AD4- libnfsidmap-0.25-4.x86+AF8-64
> > +AD4- nfs-utils-1.2.7-1.x86+AF8-64
> > +AD4- nfs-utils-clients-1.2.7-1.x86+AF8-64
> > +AD4- nfs-utils-common-1.2.7-1.x86+AF8-64
> > +AD4- 
> > +AD4- BR,
> > +AD4- Pawe+AUI-.
> > 
> > It's hard to guess when you don't tell us anything about what you are
> > doing to trigger this message, but it sounds as if you are trying to use
> > the cachefs mount options incorrectly.
> 
> on the server side there're following exports:
> 
> # exportfs -v
> /home/atest/farm-local
>                 <world>(rw,async,wdelay,nohide,crossmnt,no_root_squash,no_subtree_check,insecure_locks)
> /home/atest/farm
>                 <world>(rw,async,wdelay,nohide,crossmnt,no_root_squash,no_subtree_check,insecure_locks)
> /home/users     <world>(rw,no_root_squash,no_subtree_check,insecure_locks)
> 
> in the exported /home/atest/* there're crossmounted/binded directories:
> 
> # fstab
> /home/atest/farm-local/results  /home/atest/farm/results        none bind
> /home/services/ftp              /var/lib/ftp                    none bind
> /home/services/ftp              /home/atest/farm-local/ftp      none bind
> 
> # df -h
> Filesystem                   Size  Used Avail Use% Mounted on
> rootfs                       7.6G  4.5G  2.7G  63% /
> /dev/md0                     7.6G  4.5G  2.7G  63% /
> run                           32G  264K   32G   1% /run
> devtmpfs                      32G     0   32G   0% /dev
> /dev/md1                     7.2T  794G  6.4T  11% /home
> none                          32G     0   32G   0% /dev/shm
> /dev/md1                     7.2T  794G  6.4T  11% /home/atest/farm/results
> /dev/md1                     7.2T  794G  6.4T  11% /var/lib/ftp
> /dev/md1                     7.2T  794G  6.4T  11% /home/atest/farm-local/ftp
> 
> 
> on the client side there's only a simple autofs configuration:
> 
> # cat /etc/autofs/auto.server
> farm           -rw,soft,rsize=32768,wsize=32768,intr,vers=4,async,bg,noatime,retry=30,retrans=10        192.168.2.100:/home/atest/farm-local
> home           -rw,soft,vers=4        192.168.2.100:/home/users
> 

OK. If you are not specifying '-ofsc', then that warning is 100%
confusing. How about the following patch?

8<---------------------------------------------------------------
>From e8ed657ba5edc8c9339e00ab2fd235ab1dd94cd5 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Wed, 19 Dec 2012 16:56:28 -0500
Subject: [PATCH] NFS: Kill fscache warnings when mounting without -ofsc
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The fscache code will currently bleat a "non-unique superblock keys"
warning even if the user is mounting without the 'fsc' option.
There should be no reason to even initialise the superblock cache cookie
unless we're planning on using fscache for something, so ensure that
we check for the NFS_OPTION_FSCACHE flag before calling into the
fscache code.

Reported-by: Paweł Sikora <pawel.sikora@agmk.net>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: David Howells <dhowells@redhat.com>
---
 fs/nfs/super.c | 19 +++++++++++++++----
 1 file changed, 15 insertions(+), 4 deletions(-)

diff --git a/fs/nfs/super.c b/fs/nfs/super.c
index aa5315b..c25cadf8 100644
--- a/fs/nfs/super.c
+++ b/fs/nfs/super.c
@@ -2375,19 +2375,30 @@ static void nfs_get_cache_cookie(struct super_block *sb,
 				 struct nfs_parsed_mount_data *parsed,
 				 struct nfs_clone_mount *cloned)
 {
+	struct nfs_server *nfss = NFS_SB(sb);
 	char *uniq = NULL;
 	int ulen = 0;
 
-	if (parsed && parsed->fscache_uniq) {
-		uniq = parsed->fscache_uniq;
-		ulen = strlen(parsed->fscache_uniq);
+	nfss->fscache_key = NULL;
+	nfss->fscache = NULL;
+
+	if (parsed) {
+		if (!(parsed->options & NFS_OPTION_FSCACHE))
+			return;
+		if (parsed->fscache_uniq) {
+			uniq = parsed->fscache_uniq;
+			ulen = strlen(parsed->fscache_uniq);
+		}
 	} else if (cloned) {
 		struct nfs_server *mnt_s = NFS_SB(cloned->sb);
+		if (!(mnt_s->options & NFS_OPTION_FSCACHE))
+			return;
 		if (mnt_s->fscache_key) {
 			uniq = mnt_s->fscache_key->key.uniquifier;
 			ulen = mnt_s->fscache_key->key.uniq_len;
 		};
-	}
+	} else
+		return;
 
 	nfs_fscache_get_super_cookie(sb, uniq, ulen);
 }
-- 
1.7.11.7


-- 
Trond Myklebust
Linux NFS client maintainer

NetApp
Trond.Myklebust@netapp.com
www.netapp.com

^ permalink raw reply related	[flat|nested] 6+ messages in thread

* Re: [3.6.11] NFS: Cache request denied due to non-unique superblock keys.
  2012-12-19 22:06     ` Myklebust, Trond
@ 2012-12-20 13:42       ` David Howells
  0 siblings, 0 replies; 6+ messages in thread
From: David Howells @ 2012-12-20 13:42 UTC (permalink / raw)
  To: Myklebust, Trond
  Cc: dhowells, =?utf-7?B?UGF3ZStBVUktIFNpa29yYQ==?=,
	linux-nfs@vger.kernel.org

Myklebust, Trond <Trond.Myklebust@netapp.com> wrote:

> From e8ed657ba5edc8c9339e00ab2fd235ab1dd94cd5 Mon Sep 17 00:00:00 2001
> From: Trond Myklebust <Trond.Myklebust@netapp.com>
> Date: Wed, 19 Dec 2012 16:56:28 -0500
> Subject: [PATCH] NFS: Kill fscache warnings when mounting without -ofsc
> MIME-Version: 1.0
> Content-Type: text/plain; charset=UTF-8
> Content-Transfer-Encoding: 8bit
> 
> The fscache code will currently bleat a "non-unique superblock keys"
> warning even if the user is mounting without the 'fsc' option.
> There should be no reason to even initialise the superblock cache cookie
> unless we're planning on using fscache for something, so ensure that
> we check for the NFS_OPTION_FSCACHE flag before calling into the
> fscache code.
> 
> Reported-by: Paweł Sikora <pawel.sikora@agmk.net>
> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
> Cc: David Howells <dhowells@redhat.com>

Acked-by: David Howells <dhowells@redhat.com>

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2012-12-20 13:42 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-12-19 18:58 [3.6.11] NFS: Cache request denied due to non-unique superblock keys Paweł Sikora
2012-12-19 19:30 ` Jim Rees
2012-12-19 19:35 ` Myklebust, Trond
2012-12-19 21:16   ` Paweł Sikora
2012-12-19 22:06     ` Myklebust, Trond
2012-12-20 13:42       ` David Howells

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.