* [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.