Thanks for all the great replies.
Going back to older kernels is not an option for us, we will be looking into upgrading to lustre 2.5.x soon I hope.
In the mean time I'll probably try nfs-ganesha


On Thu, Jun 19, 2014 at 7:23 AM, Nguyen Viet Cuong <mrcuongnv@gmail.com> wrote:
Hi JC,

I wonder whether we can utilize pNFS from NFS v4.1 implemented by NFS-Ganesha for accessing Lustre in parallel way?

Regards,
Cuong


On Wed, Jun 18, 2014 at 5:38 PM, Jacques-Charles Lafoucriere <jacques-charles.lafoucriere@cea.fr> wrote:
Hello

we are using NFS-ganesha (https://github.com/nfs-ganesha/nfs-ganesha) to export Lustre.
As it is a user space NFS server, you "just" need to have a running Lustre client.

Bye

JC


On 06/16/2014 08:49 PM, E.S. Rosenberg wrote:
In the interest of easy data access from computers that are not part of the cluster we would like to export the lustre filesystem as NFS from one of the nodes.

From what I understood this should be possible but so far we are getting kernel panics.

So:
- Has anyone done it?
- What are the pitfalls?
- Any other useful tips?

Tech details:
Lustre: 2.4.3
Kernel: 3.14.3 + aufs
Distro: Debian testing/sid

We will probably be upgrading to lustre 2.5.x in the near future.

Thanks,
Eli

Trace from the test subject:
Jun 16 18:07:17 kernel:LustreError: 3795:0:(llite_internal.h:1141:ll_inode2fid()) ASSERTION( inode != ((void *)0) ) failed:
Jun 16 18:07:17 kernel:LustreError: 3795:0:(llite_internal.h:1141:ll_inode2fid()) LBUG
Jun 16 18:07:17 kernel:CPU: 0 PID: 3795 Comm: nfsd Tainted: G        WC   3.14.3-aufs-mos-1 #1
Jun 16 18:07:17 kernel:Hardware name: Dell Inc. PowerEdge C6220/03C9JJ, BIOS 1.2.1 05/27/2013
Jun 16 18:07:17 kernel: 0000000000000000 ffff881047dd5ba0 ffffffff8175c9e4 ffffffffa1842970
Jun 16 18:07:17 kernel: ffff881047dd5bc0 ffffffffa006954c 0000000000000000 ffff880845cd5148
Jun 16 18:07:17 kernel: ffff881047dd5c00 ffffffffa18054fd 0cb158b46edf5345 0000000000000013
Jun 16 18:07:17 kernel:Call Trace:
Jun 16 18:07:17 kernel: [<ffffffff8175c9e4>] dump_stack+0x45/0x56
Jun 16 18:07:17 kernel: [<ffffffffa006954c>] lbug_with_loc+0x3c/0x90 [libcfs]
Jun 16 18:07:17 kernel: [<ffffffffa18054fd>] ll_encode_fh+0x109/0x13e [lustre]
Jun 16 18:07:17 kernel: [<ffffffff81203f79>] exportfs_encode_inode_fh+0x1b/0x86
Jun 16 18:07:17 kernel: [<ffffffff8120402f>] exportfs_encode_fh+0x4b/0x60
Jun 16 18:07:17 kernel: [<ffffffff810f420f>] ? lookup_real+0x27/0x42
Jun 16 18:07:17 kernel: [<ffffffff81207689>] _fh_update.part.7+0x39/0x48
Jun 16 18:07:17 kernel: [<ffffffff81207c2a>] fh_compose+0x3d1/0x3fa
Jun 16 18:07:17 kernel: [<ffffffff81210fe4>] encode_entryplus_baggage+0xd3/0x125
Jun 16 18:07:17 kernel: [<ffffffff8121121f>] encode_entry.isra.14+0x150/0x2cb
Jun 16 18:07:17 kernel: [<ffffffff8121247d>] nfs3svc_encode_entry_plus+0xf/0x11
Jun 16 18:07:17 kernel: [<ffffffff81209e7e>] nfsd_readdir+0x160/0x1f8
Jun 16 18:07:17 kernel: [<ffffffff8121246e>] ? nfs3svc_encode_entry+0xe/0xe
Jun 16 18:07:17 kernel: [<ffffffff8120831b>] ? nfsd_splice_actor+0xe8/0xe8
Jun 16 18:07:17 kernel: [<ffffffff81056154>] ? groups_free+0x22/0x44
Jun 16 18:07:17 kernel: [<ffffffff8120fa3d>] nfsd3_proc_readdirplus+0xe3/0x1df
Jun 16 18:07:17 kernel: [<ffffffff81205269>] nfsd_dispatch+0xca/0x1ad
Jun 16 18:07:17 kernel: [<ffffffff8173579b>] svc_process+0x469/0x768
Jun 16 18:07:17 kernel: [<ffffffff81204d39>] nfsd+0xc5/0x117
Jun 16 18:07:17 kernel: [<ffffffff81204c74>] ? nfsd_destroy+0x6b/0x6b
Jun 16 18:07:17 kernel: [<ffffffff81051764>] kthread+0xd6/0xde
Jun 16 18:07:17 kernel: [<ffffffff8105168e>] ? kthread_create_on_node+0x15d/0x15d
Jun 16 18:07:17 kernel: [<ffffffff8176400c>] ret_from_fork+0x7c/0xb0
Jun 16 18:07:17 kernel: [<ffffffff8105168e>] ? kthread_create_on_node+0x15d/0x15d


_______________________________________________
HPDD-discuss mailing list
HPDD-discuss@lists.01.org
https://lists.01.org/mailman/listinfo/hpdd-discuss


_______________________________________________
HPDD-discuss mailing list
HPDD-discuss@lists.01.org
https://lists.01.org/mailman/listinfo/hpdd-discuss




--
Nguyen Viet Cuong

_______________________________________________
HPDD-discuss mailing list
HPDD-discuss@lists.01.org
https://lists.01.org/mailman/listinfo/hpdd-discuss