Just ran into that myself this morning - in my case, it was dealt with by disabling SELinux (setting it to disabled in /etc/sysconfig/selinux). I've only been doing testing on non-production systems though, so I didn't look into what specifically SELinux was throwing a fit about/modification to SELinux configuration to fix it.<br>
<br>-- <br>--------------------------------------------------------------------------------------------------<br>Gary Gogick<br>senior systems administrator | workhabit,inc.<br><br><br><div class="gmail_quote">On Fri, Jun 19, 2009 at 3:06 PM, Bill Wichser <span dir="ltr"><<a href="mailto:bill@princeton.edu">bill@princeton.edu</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">New MDS. Kernel 2.6.18-92.1.17.el5_lustre.1.8.0smp on a base RH5.3<br>
system. Nehalem CPU.<br>
<br>
I'm stuck. Did I miss some step here? The filesystem is about 300G.<br>
<br>
Thanks,<br>
Bill<br>
<br>
=======================================================<br>
<br>
I make the filesystem:<br>
<br>
[root@mds ~]# mkfs.lustre --fsname=lustre --mgs --mdt /dev/sdb1<br>
<br>
Permanent disk data:<br>
Target: lustre-MDTffff<br>
Index: unassigned<br>
Lustre FS: lustre<br>
Mount type: ldiskfs<br>
Flags: 0x75<br>
(MDT MGS needs_index first_time update )<br>
Persistent mount opts: errors=remount-ro,iopen_nopriv,user_xattr<br>
Parameters: mdt.group_upcall=/usr/sbin/l_getgroups<br>
<br>
checking for existing Lustre data: not found<br>
device size = 285561MB<br>
2 6 18<br>
formatting backing filesystem ldiskfs on /dev/sdb1<br>
target name lustre-MDTffff<br>
4k blocks 0<br>
options -J size=400 -i 4096 -I 512 -q -O<br>
dir_index,uninit_groups -F<br>
mkfs_cmd = mkfs.ext2 -j -b 4096 -L lustre-MDTffff -J size=400 -i 4096<br>
-I 512 -q -O dir_index,uninit_groups -F /dev/sdb1<br>
Writing CONFIGS/mountdata<br>
<br>
===================================================<br>
Then try to mount it:<br>
<br>
[root@mds ~]# mount -t lustre /dev/sdb1 /MDT<br>
mount.lustre: mount /dev/sdb1 at /MDT failed: Operation not supported<br>
<br>
====================================================<br>
<br>
The errors in /var/log/messages are:<br>
Jun 19 14:01:25 mds kernel: kjournald starting. Commit interval 5 seconds<br>
Jun 19 14:01:25 mds kernel: LDISKFS FS on sdb1, internal journal<br>
Jun 19 14:01:25 mds kernel: LDISKFS-fs: mounted filesystem with ordered<br>
data mode.<br>
Jun 19 14:01:25 mds kernel: kjournald starting. Commit interval 5 seconds<br>
Jun 19 14:01:25 mds kernel: LDISKFS FS on sdb1, internal journal<br>
Jun 19 14:01:25 mds kernel: LDISKFS-fs: mounted filesystem with ordered<br>
data mode.<br>
Jun 19 14:01:25 mds kernel: Lustre: MGS MGS started<br>
Jun 19 14:01:25 mds kernel: Lustre: MGC172.23.10.4@tcp: Reactivating import<br>
Jun 19 14:01:25 mds kernel: Lustre: Setting parameter<br>
lustre-MDT0000.mdt.group_upcall in log lustre-MDT0000<br>
Jun 19 14:01:25 mds kernel: Lustre: Enabling user_xattr<br>
Jun 19 14:01:25 mds kernel: Lustre: lustre-MDT0000: new disk, initializing<br>
Jun 19 14:01:25 mds kernel: Lustre: MDT lustre-MDT0000 now serving<br>
lustre-MDT0000_UUID<br>
(lustre-MDT0000/dc93cffc-68e5-8351-46eb-52215bd7a771) with recovery enabled<br>
Jun 19 14:01:25 mds kernel: Lustre:<br>
3177:0:(lproc_mds.c:271:lprocfs_wr_group_upcall()) lustre-MDT0000: group<br>
upcall set to /usr/sbin/l_getgroups<br>
Jun 19 14:01:25 mds kernel: Lustre: lustre-MDT0000.mdt: set parameter<br>
group_upcall=/usr/sbin/l_getgroups<br>
Jun 19 14:01:25 mds kernel: Lustre: Server lustre-MDT0000 on device<br>
/dev/sdb1 has started<br>
Jun 19 14:01:25 mds kernel: SELinux: (dev lustre, type lustre) has no<br>
xattr support<br>
Jun 19 14:01:25 mds kernel: Lustre: Failing over lustre-MDT0000<br>
Jun 19 14:01:25 mds kernel: Lustre: Skipped 1 previous similar message<br>
Jun 19 14:01:25 mds kernel: Lustre: *** setting obd lustre-MDT0000<br>
device 'sdb1' read-only ***<br>
Jun 19 14:01:25 mds kernel: Turning device sdb (0x800011) read-only<br>
Jun 19 14:01:25 mds kernel: Lustre: Failing over lustre-mdtlov<br>
Jun 19 14:01:25 mds kernel: Lustre: lustre-MDT0000: shutting down for<br>
failover; client state will be preserved.<br>
Jun 19 14:01:25 mds kernel: Lustre: MDT lustre-MDT0000 has stopped.<br>
Jun 19 14:01:25 mds kernel: LustreError:<br>
3072:0:(ldlm_request.c:1043:ldlm_cli_cancel_req()) Got rc -108 from<br>
cancel RPC: canceling anyway<br>
Jun 19 14:01:25 mds kernel: LustreError:<br>
3072:0:(ldlm_request.c:1632:ldlm_cli_cancel_list())<br>
ldlm_cli_cancel_list: -108<br>
Jun 19 14:01:25 mds kernel: Lustre: MGS has stopped.<br>
Jun 19 14:01:25 mds kernel: Removing read-only on unknown block (0x800011)<br>
Jun 19 14:01:25 mds kernel: Lustre: server umount lustre-MDT0000 complete<br>
<br>
[root@mds ~]# lctl network up<br>
LNET configured<br>
[root@mds ~]# lctl list_nids<br>
172.23.10.4@tcp<br>
_______________________________________________<br>
Lustre-discuss mailing list<br>
<a href="mailto:Lustre-discuss@lists.lustre.org">Lustre-discuss@lists.lustre.org</a><br>
<a href="http://lists.lustre.org/mailman/listinfo/lustre-discuss" target="_blank">http://lists.lustre.org/mailman/listinfo/lustre-discuss</a><br>
</blockquote></div><br><br>