If your other single ZFS shares are working, then I think the answer is
that the Linux client version doesn't support the nested access feature,
I'm guessing.
You could also test the nested access between your Solaris 10 10/09
server and a Solaris 10 10/09 client, if possible, to be sure this is a
Linux client issue, and not a different configuration problem.
Cindy
On 06/03/10 13:50, Cassandra Pugh wrote:
No usernames is not an issue. I have many shares that work, but they
are single zfs file systems.
The special case here is that I am trying to traverse NESTED zfs
systems, for the purpose of having compressed and uncompressed
directories.
-
Cassandra
(609) 243-2413
Unix Administrator
"From a little spark may burst a mighty flame."
-Dante Alighieri
On Thu, Jun 3, 2010 at 3:00 PM, Cindy Swearingen
<cindy.swearin...@oracle.com <mailto:cindy.swearin...@oracle.com>> wrote:
Hi Cassandra,
The mirror mount feature allows the client to access files and dirs
that are newly created on the server, but this doesn't look like
your problem
described below.
My guess is that you need to resolve the username/permission issues
before this will work, but some versions of Linux don't support
traversing nested mount points.
I'm no NFS expert and many on this list are, but things to check are:
- I'll assume that hostnames are resolving between systems since
you can share/mount the resources.
- If you are seeing "nobody" instead of user names, then you need to
make sure the domain name is specified in NFSMAPID_DOMAIN. For example,
add company.com <http://company.com> to the /etc/default/nfs file
and then restart this
server:
# svcs | grep mapid
online May_27 svc:/network/nfs/mapid:default
# svcadm restart svc:/network/nfs/mapid:default
- Permissions won't resolve correctly until the above two issues are
cleared.
- You might be able to rule out the Linux client support of nested
mount points by just sharing a simple test dataset, like this:
# zfs create mypool/test
# cp /usr/dict/words /mypool/test/file.1
# zfs set sharenfs=on mypool/test
and see if file.1 is visible on the Linux client.
Thanks,
Cindy
On 06/03/10 11:53, Cassandra Pugh wrote:
Thanks for getting back to me!
I am using Solaris 10 10/09 (update 8)
I have created multiple nested zfs directories in order to
compress some but not all sub directories in a directory.
I have ensured that they all have a sharenfs option, as I have
done with other shares.
This is a special case to me, since instead of just
#zfs create pool/mydir
and then just using mkdir to make everything thereafter, I have
done:
#zfs create mypool/mydir/
#zfs create mypool/mydir/dir1
#zfs create mypool/mydir/dir1/compressed1
#zfs create mypool/mydir/dir1/compressedir2
#zfs create mypool/mydir/dir1/uncompressedir
i had hoped that i would then export this, and mount it on the
client and see:
#ls /mnt/mydir/*
dir:
compressedir1 compressedir2 uncompressedir
and the files thereafter.
however what i see is :
#ls /mnt/mydir/*
dir:
My client is linux. I would assume we are using nfs v3. I also
notice that the permissions are not showing through correctly.
The mount options used are our "defaults"
(hard,rw,nosuid,nodev,intr,noacl)
I am not sure what this mirror mounting is? Would that help me?
Is there something else I could be doing to approach this better?
Thank you for your insight.
-
Cassandra
Unix Administrator
On Thu, May 27, 2010 at 5:25 PM, Cindy Swearingen
<cindy.swearin...@oracle.com
<mailto:cindy.swearin...@oracle.com>
<mailto:cindy.swearin...@oracle.com
<mailto:cindy.swearin...@oracle.com>>> wrote:
Cassandra,
Which Solaris release is this?
This is working for me between an Solaris 10 server and a
OpenSolaris client.
Nested mount points can be tricky and I'm not sure if you are
looking
for the mirror mount feature that is not available in the
Solaris 10
release, where new directory contents are accessible on the
client.
See the examples below.
Thanks,
Cindy
On the server:
# zpool create pool c1t3d0
# zfs create pool/myfs1
# cp /usr/dict/words /pool/myfs1/file.1
# zfs create -o mountpoint=/pool/myfs1/myfs2 pool/myfs2
# ls /pool/myfs1
file.1 myfs2
# cp /usr/dict/words /pool/myfs1/myfs2/file.2
# ls /pool/myfs1/myfs2/
file.2
# zfs set sharenfs=on pool/myfs1
# zfs set sharenfs=on pool/myfs2
# share
- /pool/myfs1 rw ""
- /pool/myfs1/myfs2 rw "
On the client:
# ls /net/t2k-brm-03/pool/myfs1
file.1 myfs2
# ls /net/t2k-brm-03/pool/myfs1/myfs2
file.2
# mount -F nfs t2k-brm-03:/pool/myfs1 /mnt
# ls /mnt
file.1 myfs2
# ls /mnt/myfs2
file.2
On the server:
# touch /pool/myfs1/myfs2/file.3
On the client:
# ls /mnt/myfs2
file.2 file.3
On 05/27/10 14:02, Cassandra Pugh wrote:
I was wondering if there is a special option to share
out a
set of nested
directories? Currently if I share out a directory with
/pool/mydir1/mydir2
on a system, mydir1 shows up, and I can see mydir2, but
nothing in mydir2.
mydir1 and mydir2 are each a zfs filesystem, each
shared with
the proper
sharenfs permissions.
Did I miss a browse or traverse option somewhere?
-
Cassandra
Unix Administrator
"From a little spark may burst a mighty flame."
-Dante Alighieri
------------------------------------------------------------------------
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
<mailto:zfs-discuss@opensolaris.org>
<mailto:zfs-discuss@opensolaris.org
<mailto:zfs-discuss@opensolaris.org>>
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
