On Tue, 18 Oct 2005, Rainer Krienke wrote:
> Hello,
>
> using the autofs package from suse10.0 (autofs-4.1.4-6.i586.rpm; autofs4
> kernel module, Kernel2.6.13-15) NFS mounts of user directories fail when
> using indirect mounts.
> The problem can be reproduced in a certrain setup describes below. Because of
> this, the description
> is a bit lenghty.
It's good to see that SuSE is attempting to keep up with the autofs
versions. If you have the oppertunity, Rainer, please thank them for me.
>
> The maps for automount are taken from NIS. To mount a users directory via
> NFS eg. /home/krienke on a
> client entries from the NIS-maps auto.home and auto.vol are used.
> auto.master is
> configured like this:
>
> /home auto.home -intr
> /import auto.vol -intr
>
> In auto.home there is this entry for each user like "krienke":
>
> krienke localhost:/import/user2/krienke
>
> In auto.vol you find another entry to handle the "user2" key starting at
> /import from above
> (there are more than one nfsservers exporting user1, user2, ... each with
> about 2000 users):
>
> user2 nfssrv2:/export/user2
>
> nfssrv2 ist one of our NFS servers for home directories. So basically we nfs
> mount a volume containing
> all the users on this server on /import/user2 on the client and then we mount
> /home/krienke on
> /import/user2/krienke on the client (the advantage of this indirect 2 layer
> mount is that we
> avoid problems with too many mounts for users on one client since bind mounts
> are used for the
> users home). This mount was (up to 9.3) automatically mounted as a bind mount
> by automount.
> Up to suse9.3 (using autofs4-4.1.3) this worked without problems and if I
> replace the suse10
> autofs rpm by the suse9.3 autofs4 rpm it again does work. I can also manually
> issue the mount
> command logged by automunt
> below (mount -t nfs -s -o intr nfssrv2:/export/user2 /import/user2) and it
> works without any problems.
As Jeff has said this sort of nesting of mounts is, in general, not
supported by autofs and it worked previously because the locking was less
broken in 4.1.3 than in 4.1.4 (yes its still broken in 4.1.4).
Can this case be done safely?
I'm not entirely sure of the answer to this but I've never liked using
locking in autofs. I introduced it only to prevent mtab corruption during
mount (correct me if you have a different recollection Jeff). It happens
during rapid mount activity such as when there as a largish number of
entries in the master map (of the order of 50 or more).
The mtab locking should be done in mount which narrows the scope of the
lock enough to allow this case to function as requested.
However, lets not forget that, even if mount were working OK today, we
would need to carry a patch in the autofs distribution for older versions.
I can see the mailing list filling with support queries already.
Can we eliminate the locking without reintroducing the problems that lead
to it being used?
Maybe, if mounts locking is working as intended but see above.
Can this be done another way?
Probably with a multi-mount map but this would introduce a fair bit of
added complexity to the map. Or possibly with a program map as Jim
has suggested.
>
> When trying to mount a user like "krienke" saying eg ls ~krienke the
> following log appears
> in /var/log/messages (autmount is running with -d -v):
>
> Oct 18 12:41:59 rzinstal1 automount[16770]: handle_packet: type = 0
> Oct 18 12:41:59 rzinstal1 automount[16770]: handle_packet_missing: token 49,
> name krienke
> Oct 18 12:41:59 rzinstal1 automount[16770]: attempting to mount entry
> /home/krienke
> Oct 18 12:41:59 rzinstal1 automount[16791]: lookup(yp): looking up krienke
> Oct 18 12:41:59 rzinstal1 automount[16791]: ret = 1
> Oct 18 12:41:59 rzinstal1 automount[16791]: lookup(yp): krienke ->
> localhost:/import/user2/krienke
> Oct 18 12:41:59 rzinstal1 automount[16791]: parse(sun): expanded entry:
> localhost:/import/user2/krienke
> Oct 18 12:41:59 rzinstal1 automount[16791]: parse(sun): gathered options:
> intr
> Oct 18 12:41:59 rzinstal1 automount[16791]: parse(sun):
> dequote("localhost:/import/user2/krienke") -> localhost:/import/user2/krienke
> Oct 18 12:41:59 rzinstal1 automount[16791]: parse(sun): core of entry:
> options=intr, loc=localhost:/import/user2/krienke
> Oct 18 12:41:59 rzinstal1 automount[16791]: parse(sun): mounting root /home,
> mountpoint krienke, what localhost:/import/user2/krienke, fstype nfs,
> Oct 18 12:41:59 rzinstal1 automount[16791]: mount(nfs): root=/home
> name=krienke what=localhost:/import/user2/krienke, fstype=nfs, options=intr
> Oct 18 12:41:59 rzinstal1 automount[16791]: mount(nfs): nfs options="intr",
> nosymlink=0, ro=0
> Oct 18 12:41:59 rzinstal1 automount[16791]: mount(nfs): winner = localhost
> local = 1
> Oct 18 12:41:59 rzinstal1 automount[16791]: mount(nfs): from
> localhost:/import/user2/krienke elected /import/user2/krienke
> Oct 18 12:41:59 rzinstal1 automount[16791]: mount(nfs): krienke is local,
> doing bind
> Oct 18 12:41:59 rzinstal1 automount[16791]: mount(bind): calling mkdir_path
> /home/krienke
> Oct 18 12:41:59 rzinstal1 automount[16791]: mount(bind): calling mount --bind
> -s -o defaults /import/user2/krienke /home/krienke
> Oct 18 12:41:59 rzinstal1 automount[16772]: handle_packet: type = 0
> Oct 18 12:41:59 rzinstal1 automount[16772]: handle_packet_missing: token 50,
> name user2
> Oct 18 12:41:59 rzinstal1 automount[16772]: attempting to mount entry
> /import/user2
> Oct 18 12:41:59 rzinstal1 automount[16793]: lookup(yp): looking up user2
> Oct 18 12:41:59 rzinstal1 automount[16793]: ret = 1
> Oct 18 12:41:59 rzinstal1 automount[16793]: lookup(yp): user2 ->
> nfssrv2:/export/user2
> Oct 18 12:41:59 rzinstal1 automount[16793]: parse(sun): expanded entry:
> nfssrv2:/export/user2
> Oct 18 12:41:59 rzinstal1 automount[16793]: parse(sun): gathered options:
> intr
> Oct 18 12:41:59 rzinstal1 automount[16793]: parse(sun):
> dequote("nfssrv2:/export/user2") -> nfssrv2:/export/user2
> Oct 18 12:41:59 rzinstal1 automount[16793]: parse(sun): core of entry:
> options=intr, loc=nfssrv2:/export/user2
> Oct 18 12:41:59 rzinstal1 automount[16793]: parse(sun): mounting root
> /import, mountpoint user2, what nfssrv2:/export/user2, fstype nfs, options
> intr
> Oct 18 12:41:59 rzinstal1 automount[16793]: mount(nfs): root=/import
> name=user2 what=nfssrv2:/export/user2, fstype=nfs, options=intr
> Oct 18 12:41:59 rzinstal1 automount[16793]: mount(nfs): nfs options="intr",
> nosymlink=0, ro=0
> Oct 18 12:41:59 rzinstal1 automount[16793]: mount(nfs): winner = nfssrv2
> local = 0
> Oct 18 12:41:59 rzinstal1 automount[16793]: mount(nfs): from
> nfssrv2:/export/user2 elected nfssrv2:/export/user2
> Oct 18 12:41:59 rzinstal1 automount[16793]: mount(nfs): calling mkdir_path
> /import/user2
> Oct 18 12:41:59 rzinstal1 automount[16793]: mount(nfs): calling mount -t nfs
> -s -o intr nfssrv2:/export/user2 /import/user2
> Oct 18 12:42:30 rzinstal1 automount[16793]: aquire_lock: can't lock lock file
> timed out: /var/lock/autofs
> Oct 18 12:42:30 rzinstal1 automount[16793]: mount(nfs): nfs: mount failure
> nfssrv2:/export/user2 on /import/user2
> Oct 18 12:42:30 rzinstal1 automount[16793]: failed to mount /import/user2
> Oct 18 12:42:30 rzinstal1 automount[16793]: umount_multi: path=/import/user2
> incl=1
> Oct 18 12:42:30 rzinstal1 automount[16793]: umount_multi: no mounts found
> under /import/user2
> Oct 18 12:42:30 rzinstal1 automount[16772]: handle_child: got pid 16793, sig
> 0 (0), stat 1
> Oct 18 12:42:30 rzinstal1 automount[16772]: sig_child: found pending iop pid
> 16793: signalled 0 (sig 0), exit status 1
> Oct 18 12:42:30 rzinstal1 automount[16772]: send_fail: token=50
>
> Then after 2 more retries to mount /import/user2 the whole mount fails.
>
> Oct 18 12:44:04 rzinstal1 automount[16798]: aquire_lock: can't lock lock file
> timed out: /var/lock/autofs
> Oct 18 12:44:04 rzinstal1 automount[16798]: mount(nfs): nfs: mount failure
> nfssrv2:/export/user2 on /import/user2
> Oct 18 12:44:04 rzinstal1 automount[16798]: failed to mount /import/user2
> Oct 18 12:44:04 rzinstal1 automount[16798]: umount_multi: path=/import/user2
> incl=1
> Oct 18 12:44:04 rzinstal1 automount[16798]: umount_multi: no mounts found
> under /import/user2
> Oct 18 12:44:04 rzinstal1 automount[16772]: handle_child: got pid 16798, sig
> 0 (0), stat 1
> Oct 18 12:44:04 rzinstal1 automount[16772]: sig_child: found pending iop pid
> 16798: signalled 0 (sig 0), exit status 1
> Oct 18 12:44:04 rzinstal1 automount[16772]: send_fail: token=54
> Oct 18 12:44:04 rzinstal1 automount[16795]: >> mount: special device
> /import/user2/krienke does not exist
> Oct 18 12:44:04 rzinstal1 automount[16795]: failed to mount /home/krienke
> Oct 18 12:44:04 rzinstal1 automount[16795]: umount_multi: path=/home/krienke
> incl=1
> Oct 18 12:44:04 rzinstal1 automount[16795]: umount_multi: no mounts found
> under /home/krienke
> Oct 18 12:44:04 rzinstal1 automount[16770]: handle_child: got pid 16795, sig
> 0 (0), stat 1
> Oct 18 12:44:04 rzinstal1 automount[16770]: sig_child: found pending iop pid
> 16795: signalled 0 (sig 0), exit status 1
> Oct 18 12:44:04 rzinstal1 automount[16770]: send_fail: token=52
>
>
> Any idea why or how to fix this?
>
> Thanks
> Rainer
> --
> ---------------------------------------------------------------------------
> Rainer Krienke, Universitaet Koblenz, Rechenzentrum, Raum A022
> Universitaetsstrasse 1, 56070 Koblenz, Tel: +49 261287 -1312, Fax: -1001312
> Mail: [EMAIL PROTECTED], Web: http://www.uni-koblenz.de/~krienke
> Get my public PGP key: http://www.uni-koblenz.de/~krienke/mypgp.html
> ---------------------------------------------------------------------------
>
> _______________________________________________
> autofs mailing list
> [email protected]
> http://linux.kernel.org/mailman/listinfo/autofs
>
_______________________________________________
autofs mailing list
[email protected]
http://linux.kernel.org/mailman/listinfo/autofs
