Lustre discuss - Oct 2006 - Unrecognized mount option "iopen_nopriv" or missing value: Lustre 1.6.0 b5

Hello,

    I''m evaluating lustre 1.6.0 b5 in a NPACI Rocks 4.2 cluster, which 
is RHEL4u4 base with kernel 2.6.9-34.0.2.ELsmp. I choose 1.6.0b5 since 
it has many appealing feature such as patch-less client and NFS export 
support. I tried to test a single-node configuration but failed. Details 
are below.

Lustre 1.6.0b5 configured with ./configure 
--with-linux=/usr/src/linux-2.6 with patched from 
https://mail.clusterfs.com/pipermail/lustre-discuss/2006-August/001932.html. 
The patch-less configuration was detected automatically by configure script.

My /etc/modprobe.conf:

alias scsi_hostadapter mptbase
alias scsi_hostadapter1 mptscsih
alias scsi_hostadapter2 ata_piix
alias scsi_hostadapter3 ips
alias usb-controller ehci-hcd
alias usb-controller1 uhci-hcd
#
# Do NOT Edit (generated by dbreport)
#
alias eth0
alias eth1 None
# start lustre config
# Lustre module options added automatically by lc_modprobe.sh
options lnet networks=tcp
# end lustre config


My lustre.csv:

# combo mdt/mgs
araya.local,options lnet 
networks=tcp,/tmp/mgs,/mnt/mgs,mgs|mdt,,,,--device-size=10240 
--backfstype=ext3

# ost0
araya.local,options lnet 
networks=tcp,/tmp/ost0,/mnt/ost0,ost,,araya.local@tcp0,,--device-size=10240 
--backfstype=ext3


    Below are output of lustre_config.sh command

[root@araya ~]# lustre_config.sh lustre.csv
lustre_config.sh: Checking the cluster network connectivity and hostnames...
lustre_config.sh: Check the cluster network connectivity and hostnames OK!

lustre_config.sh: ******** Lustre cluster configuration START ********

   Permanent disk data:
Target:     lustre-MDTffff
Index:      unassigned
Lustre FS:  lustre
Mount type: ext3
Flags:      0x75
              (MDT MGS needs_index first_time update )
Persistent mount opts: errors=remount-ro,iopen_nopriv,user_xattr
Parameters:

checking for existing Lustre data
formatting backing filesystem ext3 on /dev/loop0
        target name  lustre-MDTffff
        4k blocks     2560
        options        -i 4096 -I 512 -q -O dir_index -F
mkfs_cmd = mkfs.ext2 -j -b 4096 -L lustre-MDTffff  -i 4096 -I 512 -q -O 
dir_index -F /dev/loop0 2560
Writing CONFIGS/mountdata

   Permanent disk data:
Target:     lustre-OSTffff
Index:      unassigned
Lustre FS:  lustre
Mount type: ext3
Flags:      0x72
              (OST needs_index first_time update )
Persistent mount opts: errors=remount-ro
Parameters: mgsnode=10.1.1.1@tcp

checking for existing Lustre data
formatting backing filesystem ext3 on /dev/loop0
        target name  lustre-OSTffff
        4k blocks     2560
        options        -I 256 -q -O dir_index -F
mkfs_cmd = mkfs.ext2 -j -b 4096 -L lustre-OSTffff  -I 256 -q -O 
dir_index -F /dev/loop0 2560
Writing CONFIGS/mountdata
lustre_config.sh: ******** Lustre cluster configuration END **********

[root@araya ~]# mount -a
mount.lustre: mount /dev/loop0 at /mnt/mgs failed: Invalid argument
This may have multiple causes.
Are the mount options correct?
Check the syslog for more info.
mount.lustre: mount /dev/loop0 at /mnt/ost0 failed: No such device
Are the lustre modules loaded?
Check /etc/modprobe.conf and /proc/filesystems
Note ''alias lustre llite'' should be removed from modprobe.conf

[root@araya ~]# dmesg
EXT3 FS on loop0, internal journal
EXT3-fs: mounted filesystem with ordered data mode.
kjournald starting.  Commit interval 5 seconds
EXT3 FS on loop0, internal journal
EXT3-fs: mounted filesystem with ordered data mode.
Lustre:   mount data:
Lustre: device:  /dev/loop0
Lustre: flags:   0
LustreError: 9552:0:(obd_mount.c:1213:server_kernel_mount()) premount 
ldiskfs failed (-19), trying ext3
kjournald starting.  Commit interval 5 seconds
EXT3 FS on loop0, internal journal
EXT3-fs: mounted filesystem with ordered data mode.
Lustre:   disk data:
Lustre: server:  lustre-MDTffff
Lustre: uuid:   
Lustre: fs:      lustre
Lustre: index:   ffff
Lustre: config:  1
Lustre: flags:   0x75
Lustre: diskfs:  ext3
Lustre: options: errors=remount-ro,iopen_nopriv,user_xattr
Lustre: params: 
Lustre: comment:
EXT3-fs: Unrecognized mount option "iopen_nopriv" or missing value
LustreError: 9552:0:(obd_mount.c:1266:server_kernel_mount()) 
do_kern_mount failed: rc = -22
LustreError: 9552:0:(obd_mount.c:1478:server_fill_super()) Unable to 
mount device /dev/loop0: -22
LustreError: 9552:0:(obd_mount.c:1857:lustre_fill_super()) Unable to 
mount  (-22)
Lustre:   mount data:
Lustre: device:  /dev/loop0
Lustre: flags:   0
LustreError: 9558:0:(obd_mount.c:1213:server_kernel_mount()) premount 
ldiskfs failed (-19), trying ext3
kjournald starting.  Commit interval 5 seconds
EXT3 FS on loop0, internal journal
EXT3-fs: mounted filesystem with ordered data mode.
Lustre:   disk data:
Lustre: server:  lustre-OSTffff
Lustre: uuid:   
Lustre: fs:      lustre
Lustre: index:   ffff
Lustre: config:  1
Lustre: flags:   0x72
Lustre: diskfs:  ext3
Lustre: options: errors=remount-ro
Lustre: params:   mgsnode=10.1.1.1@tcp
Lustre: comment:
kjournald starting.  Commit interval 5 seconds
EXT3 FS on loop0, internal journal
EXT3-fs: mounted filesystem with ordered data mode.
Lustre: 9570:0:(lib-move.c:1624:lnet_parse_put()) Dropping PUT from 
12345-0@lo portal 26 match 1 offset 0 length 240: 2
LustreError: Can''t load module ''ost''
LustreError: 9558:0:(genops.c:223:class_newdev()) OBD: unknown type: ost
LustreError: 9558:0:(obd_config.c:175:class_attach()) Cannot create 
device OSS of type ost : -19
LustreError: 9558:0:(obd_mount.c:445:lustre_start_simple()) OSS attach 
error -19
LustreError: 9558:0:(obd_mount.c:1031:server_start_targets()) failed to 
start OSS: -19
LustreError: 9558:0:(obd_mount.c:1514:server_fill_super()) Unable to 
start targets: -19
LustreError: 9558:0:(mgc_request.c:147:config_log_find()) can''t get log
lustre-OSTffff
LustreError: 9558:0:(obd_mount.c:1342:server_put_super()) no obd 
lustre-OSTffff
LustreError: 9558:0:(obd_mount.c:119:server_deregister_mount()) 
lustre-OSTffff not registered
Lustre: server umount lustre-OSTffff complete
LustreError: 9558:0:(obd_mount.c:1857:lustre_fill_super()) Unable to 
mount  (-19)


    Does "Unrecognized mount option "iopen_nopriv" or missing
value"
normal? I suppose this should be the cause of kern_mount error. Is this 
because of patchless kernel?
    I tried to look into this mailing-list and found no clues. Sorry if 
the question is too stupid.


-- 
-----------------------------------------------------------------------------------
Somsak Sriprayoonsakul

Thai National Grid Center
Software Industry Promotion Agency
Ministry of ICT, Thailand
somsak_sr@thaigrid.or.th
-----------------------------------------------------------------------------------

On Oct 05, 2006  01:18 +0700, Somsak Sriprayoonsakul
wrote:
> LustreError: 9552:0:(obd_mount.c:1213:server_kernel_mount()) premount 
> ldiskfs failed (-19), trying ext3
>
> EXT3-fs: Unrecognized mount option "iopen_nopriv" or missing
value
> LustreError: 9552:0:(obd_mount.c:1266:server_kernel_mount()) 
> do_kern_mount failed: rc = -22
> LustreError: 9552:0:(obd_mount.c:1478:server_fill_super()) Unable to 
> mount device /dev/loop0: -22
> LustreError: 9552:0:(obd_mount.c:1857:lustre_fill_super()) Unable to 
> mount  (-22)
This is both something wrong in your environment (somehow "ldiskfs.ko"
is not found), and in the mount program (which shouldn''t be trying ext3
unless it is running on a 2.4 kernel, produces a confusing error message
instead of "ldiskfs could not be found" or similar).

Cheers, Andreas
--
Andreas Dilger
Principal Software Engineer
Cluster File Systems, Inc.

An HTML attachment was scrubbed...
URL:
http://mail.clusterfs.com/pipermail/lustre-discuss/attachments/20061004/299e992d/attachment.html

On Oct 05, 2006  10:26 +0700, Somsak Sriprayoonsakul
wrote:
> It use ext3 because I force it in lustre.csv (--backfstype=ext3). I
> need to force ext3 because ldiskfs module is not built. I check in the
> configure script and found that ldiskfs will not be built by default if
> patchless client is possible, so I assumed that I should use ext3.
> Seems that I need to patch the kernel for servers isn''t it?
The patchless CLIENT does not need patches.  The servers need to be
patched in order to run Lustre.  You need to get a separate kernel
build and lustre build for the patched kernel.  You can download these
from CFS or build your own.

Cheers, Andreas
--
Andreas Dilger
Principal Software Engineer
Cluster File Systems, Inc.

An HTML attachment was scrubbed...
URL:
http://mail.clusterfs.com/pipermail/lustre-discuss/attachments/20061005/f90b9fae/attachment.html

On Oct 05, 2006  15:13 +0700, Somsak Sriprayoonsakul
wrote:
> Is it possible to use 1.4.7 patched kernel?
Yes, that should be fine.
> If not, which patch set do I need for the kernel?
> ./2.6-rhel4.series
This should be applied to the 2.6.9-42.EL RHEL4 kernel.
> ./ldiskfs-2.6-rhel4.series
This is applied automatically during the lustre build to create the
ldiskfs module, it should not be applied by you to the linux kernel.

Cheers, Andreas
--
Andreas Dilger
Principal Software Engineer
Cluster File Systems, Inc.