Thanks - tunefs worked fine on MDS/MGT combo server.
On OSS tunefs fails with "unsupported features" message. Both MDS/MGT
and OSS are running 2.6.9-42 kernel.
I'd expect tunefs to fail on both servers if the kernel is to old. I'm
using Lustre 1.5.95.
Tim
On OSS:
[EMAIL PROTECTED] ~]# umount /dev/sdc1
[EMAIL PROTECTED] ~]# tunefs.lustre --writeconf /dev/sdc1
checking for existing Lustre data
/dev/sdc1: Filesystem has unsupported feature(s) while opening
filesystem
In all likelihood, the 'unsupported feature' is 'extents', which
older debugfs does not understand.
Use e2fsprogs-1.38-cfs1 or later, available from
ftp://ftp.lustre.org/pub/lustre/other/e2fsprogs/
found Lustre data
tunefs.lustre: Unable to read CONFIGS/mountdata (No such file or
directory).
Contents of CONFIGS:
Trying last_rcvd
tunefs.lustre: Unable to read old data
tunefs.lustre FATAL: Failed to read previous Lustre data from
/dev/sdc1
[EMAIL PROTECTED] ~]# uname -a
Linux Redhat166 2.6.9-42.EL_lustre.1.5.95smp #1 SMP Thu Sep 28
06:36:13 MDT 2006 i686 i686 i386 GNU/Linux
[EMAIL PROTECTED] ~]# tunefs.lustre -h
tunefs.lustre v1.5.95
usage: tunefs.lustre <target types> [options] <device>
On MDS /MGT:
[EMAIL PROTECTED] ~]# umount /dev/sdb1
[EMAIL PROTECTED] ~]# tunefs.lustre --writeconf /dev/sdb1
checking for existing Lustre data
found Lustre data
Reading CONFIGS/mountdata
Read previous values:
Target: test-MDT0000
Index: 0
Lustre FS: test
Mount type: ldiskfs
Flags: 0x5
(MDT MGS )
Persistent mount opts: errors=remount-ro,iopen_nopriv,user_xattr
Parameters:
Permanent disk data:
Target: test-MDT0000
Index: 0
Lustre FS: test
Mount type: ldiskfs
Flags: 0x105
(MDT MGS writeconf )
Persistent mount opts: errors=remount-ro,iopen_nopriv,user_xattr
Parameters:
Writing CONFIGS/mountdata
[EMAIL PROTECTED] ~]# uname -a
Linux Redhat101 2.6.9-42.EL_lustre.1.5.95smp #1 SMP Thu Sep 28
06:36:13 MDT 2006 i686 i686 i386 GNU/Linux
-----Original Message-----
From: Nathaniel Rutman [mailto:[EMAIL PROTECTED]
Sent: Friday, August 03, 2007 7:55 PM
To: Snider, Tim
Cc: [email protected]
Subject: Re: [Lustre-discuss] Problems switching the OSS and getting
Lustre to restart correctly.
Assuming [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]> is the old OSS,
the ptlrpc_expire_one_request()) @@@ timeout messages mean that the
client / MDT was trying and failing to talk to the old server.
You need to tell Lustre to regenerate the configuration logs using
'tunefs.lustre --writeconf' -- see
http://wiki.lustre.org/index.php?title=Mount_Conf#Changing_a_server_nid
Snider, Tim wrote:
> I have a simple configuration where I'd like to switch the OSS to a
> different server. The OST is on external storage and will remain the
> same. I'll switch cables to the storage between servers. The MDS, MGT
> and client remain the same. After rebooting all machines, Lustre
> seems to start correctly again on the MDS/MGT and OSS - no console
> messages. I can also mount the client without any console errors,
> however an ls command on the client mounted device hangs.
>
> entries in /var/log/messages on the MDS indicate there was an error
> from the old OSS - which isn't involved in the Lustre configuration at
> this point:
> Lustre: 5012:0:(peer.c:238:lnet_debug_peer())
> [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]> 2 up
> 8 8 8 8 7 0
> OLD OSS = 172.22.14.245 (not in use - but still on the network)
> Current OSS IP = 172.22.14.166
> MDS/MGT = 172.22.14.101
> Client = 172.22.14.100
>
> How do you properly switch out the OSS and restart using the same
OSTs?
> Thanks
> Tim
>
> Jul 31 17:54:58 Redhat101 kernel: Lustre:
> 4808:0:(module.c:382:init_libcfs_module()) maximum lustre stack 8192
> Jul 31 17:54:58 Redhat101 kernel: Lustre: OBD class driver,
> [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>
> Jul 31 17:54:58 Redhat101 kernel: Lustre Version: 1.5.95
> Jul 31 17:54:58 Redhat101 kernel: Build Version:
> 1.5.95-19691231170000-PRISTINE-.testsuite.tmp.boulder.lbuild-boulder.B
> UILD.lustre-kernel-2.6.9.lustre.linux-2.6.9-42.EL_lustre.1.5.95smp
> Jul 31 17:54:58 Redhat101 kernel: Lustre: Added LNI [EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]> [8/256] Jul 31 17:54:58 Redhat101 kernel:
> Lustre: Accept secure, port 988 Jul 31 17:54:58 Redhat101 kernel:
> Lustre: Lustre Client File System; [EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]>
> Jul 31 17:54:58 Redhat101 kernel: Lustre: mount data:
> Jul 31 17:54:58 Redhat101 kernel: Lustre: device: /dev/sdb1
> Jul 31 17:54:58 Redhat101 kernel: Lustre: flags: 0
> Jul 31 17:54:59 Redhat101 kernel: kjournald starting. Commit interval
> 5 seconds
> Jul 31 17:54:59 Redhat101 kernel: LDISKFS FS on sdb1, internal journal
> Jul 31 17:54:59 Redhat101 kernel: LDISKFS-fs: mounted filesystem with
> ordered data mode.
> Jul 31 17:54:59 Redhat101 kernel: Lustre: disk data:
> Jul 31 17:54:59 Redhat101 kernel: Lustre: server: test-MDT0000 Jul 31
> 17:54:59 Redhat101 kernel: Lustre: uuid:
> Jul 31 17:54:59 Redhat101 kernel: Lustre: fs: test
> Jul 31 17:54:59 Redhat101 kernel: Lustre: index: 0000
> Jul 31 17:54:59 Redhat101 kernel: Lustre: config: 2
> Jul 31 17:54:59 Redhat101 kernel: Lustre: flags: 0x5
> Jul 31 17:54:59 Redhat101 kernel: Lustre: diskfs: ldiskfs Jul 31
> 17:54:59 Redhat101 kernel: Lustre: options:
> errors=remount-ro,iopen_nopriv,user_xattr
> Jul 31 17:54:59 Redhat101 kernel: Lustre: params:
> Jul 31 17:54:59 Redhat101 kernel: Lustre: comment:
> Jul 31 17:54:59 Redhat101 kernel: kjournald starting. Commit interval
> 5 seconds
> Jul 31 17:54:59 Redhat101 kernel: LDISKFS FS on sdb1, internal journal
> Jul 31 17:54:59 Redhat101 kernel: LDISKFS-fs: mounted filesystem with
> ordered data mode.
> Jul 31 17:54:59 Redhat101 kernel: Lustre: 0 UP mgs MGS MGS 5
> Jul 31 17:54:59 Redhat101 kernel: Lustre: 1 UP mgc
> [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>
> 874c230c-bc4b-f2df-7498-9680ca5495c6 6
> Jul 31 17:54:59 Redhat101 kernel: Lustre: 2 UP mdt MDS MDS_uuid 3
> Jul 31 17:54:59 Redhat101 kernel: Lustre: 3 UP lov test-mdtlov
> test-mdtlov_UUID 4
> Jul 31 17:54:59 Redhat101 kernel: Lustre: 4 UP mds test-MDT0000
> test-MDT0000_UUID 4
> Jul 31 17:54:59 Redhat101 kernel: Lustre: 5 UP osc test-OST0000-osc
> test-mdtlov_UUID 5
> Jul 31 17:55:01 Redhat101 crond(pam_unix)[5124]: session opened for
> user root by (uid=0) Jul 31 17:55:02 Redhat101 crond(pam_unix)[5124]:
> session closed for user root Jul 31 17:55:04 Redhat101 kernel: Lustre:
> 5012:0:(peer.c:238:lnet_debug_peer()) [EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]> 2 up 8 8 8
> 8 7 0
> Jul 31 17:55:29 Redhat101 kernel: LustreError:
> 5012:0:(client.c:950:ptlrpc_expire_one_request()) @@@ timeout (sent at
> 1185918924, 5s ago) Jul 31 17:55:29 Redhat101 kernel: LustreError:
> 5012:0:(client.c:950:ptlrpc_expire_one_request()) Skipped 2 previous
> similar messages Jul 31 17:55:29 Redhat101 kernel: Lustre:
> 5012:0:(peer.c:238:lnet_debug_peer()) [EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]> 2 up 8 8 8
> 8 7 0
> Jul 31 17:55:54 Redhat101 kernel: LustreError:
> 5012:0:(client.c:950:ptlrpc_expire_one_request()) @@@ timeout (sent at
> 1185918949, 5s ago) Jul 31 17:55:54 Redhat101 kernel: LustreError:
> 5012:0:(client.c:950:ptlrpc_expire_one_request()) Skipped 1 previous
> similar message Jul 31 17:55:54 Redhat101 kernel: Lustre:
> 5012:0:(peer.c:238:lnet_debug_peer()) [EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]> 2 up 8 8 8
> 8 7 0
> Jul 31 17:56:19 Redhat101 kernel: LustreError:
> 5012:0:(client.c:950:ptlrpc_expire_one_request()) @@@ timeout (sent at
> 1185918974, 5s ago) Jul 31 17:56:19 Redhat101 kernel: LustreError:
> 5012:0:(client.c:950:ptlrpc_expire_one_request()) Skipped 1 previous
> similar message Jul 31 17:56:19 Redhat101 kernel: Lustre:
> 5012:0:(peer.c:238:lnet_debug_peer()) [EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]> 2 up 8 8 8
> 8 7 0
> ----------------------------------------------------------------------
> --
>
> _______________________________________________
> Lustre-discuss mailing list
> [email protected]
> https://mail.clusterfs.com/mailman/listinfo/lustre-discuss
>
_______________________________________________
Lustre-discuss mailing list
[email protected]
https://mail.clusterfs.com/mailman/listinfo/lustre-discuss
