[Lustre-discuss] Lustre client not mount after upgrade

Andreas Dilger adilger at whamcloud.com
Wed Oct 26 01:58:30 PDT 2011


On 2011-10-24, at 10:43 PM, Ekaterina Popova wrote:
> I have upgraded Lustre client on my machine till 2.1.50 version. Since i 
> can't mount client.

Don't do that.  The "master" branch (2.1.50) is the development branch
for 2.2 and is starting to get new features landed and has not gone
through more than basic regression testing (i.e. no interoperability
testing).

I would recommend using 2.1.0 (clients & servers).

> mount -t lustre 192.168.1.10 at tcp0:/lihep /lustre
> mount.lustre: mount  192.168.1.10 at tcp0:/lihep at /lustre failed: No such 
> file or directory
> Is the MGS specification correct?
> Is the filesystem name correct?
> If upgrading, is the copied client log valid? (see upgrade docs)
> 
> At syslog:
> 
> Oct 25 08:34:00 draco kernel: [73630.215855] Lustre: Lustre: Build 
> Version: v2_1_50_0-ga39459d-CHANGED-2.6.32
> Oct 25 08:34:00 draco kernel: [73630.275393] Lustre: Added LNI 
> 192.168.1.1 at tcp [8/256/0/180]
> Oct 25 08:34:00 draco kernel: [73630.275422] Lustre: Accept secure, port 988
> Oct 25 08:34:00 draco kernel: [73630.314580] Lustre: Lustre OSC module 
> (ffffffffa08973c0).
> Oct 25 08:34:00 draco kernel: [73630.328461] Lustre: Lustre LOV module 
> (ffffffffa08f2820).
> Oct 25 08:34:00 draco kernel: [73630.347306] Lustre: Lustre client 
> module (ffffffffa0987880).
> Oct 25 08:34:00 draco kernel: [73630.356016] Lustre: import 
> MGC192.168.1.10 at tcp->MGC192.168.1.10 at tcp_0 netid 20000: select flavor null
> Oct 25 08:34:00 draco kernel: [73630.358509] Lustre: 
> MGC192.168.1.10 at tcp: Reactivating import
> Oct 25 08:34:00 draco kernel: [73630.370731] LustreError: 
> 23799:0:(ldlm_lib.c:382:client_obd_setup()) can't add initial connection
> Oct 25 08:34:00 draco kernel: [73630.370742] LustreError: 
> 23799:0:(obd_config.c:522:class_setup()) setup 
> lihep-MDT0000-mdc-ffff8801bb901000 failed (-2)
> Oct 25 08:34:00 draco kernel: [73630.370745] LustreError: 
> 23799:0:(obd_config.c:1361:class_config_llog_handler()) Err -2 on cfg 
> command:
> Oct 25 08:34:00 draco kernel: [73630.370749] Lustre:    cmd=cf003 
> 0:lihep-MDT0000-mdc  1:lihep-MDT0000_UUID  2:10.1.1.10 at tcp1
> Oct 25 08:34:00 draco kernel: [73630.370766] LustreError: 15c-8: 
> MGC192.168.1.10 at tcp: The configuration from log 'lihep-client' failed 
> (-2). This may be the result of communication errors between this node 
> and the MGS, a bad configuration, or other errors. See the syslog for 
> more information.
> Oct 25 08:34:00 draco kernel: [73630.370771] LustreError: 
> 23733:0:(llite_lib.c:949:ll_fill_super()) Unable to process log: -2
> Oct 25 08:34:00 draco kernel: [73630.370851] LustreError: 
> 23733:0:(obd_config.c:567:class_cleanup()) Device 3 not setup
> Oct 25 08:34:00 draco kernel: [73630.387299] LustreError: 
> 23733:0:(ldlm_request.c:1173:ldlm_cli_cancel_req()) Got rc -108 from 
> cancel RPC: canceling anyway
> Oct 25 08:34:00 draco kernel: [73630.387304] LustreError: 
> 23733:0:(ldlm_request.c:1800:ldlm_cli_cancel_list()) 
> ldlm_cli_cancel_list: -108
> Oct 25 08:34:00 draco kernel: [73630.387897] Lustre: client 
> ffff8801bb901000 umount complete
> Oct 25 08:34:00 draco kernel: [73630.387901] LustreError: 
> 23733:0:(obd_mount.c:2306:lustre_fill_super()) Unable to mount  (-2)
> 
> I don't upgrade MDS/OSS. Their version is 2.0.
> 
> Can I fix client? What is the reason of this error?
> 
> With regards,
> Ekaterina Popova
> 
> 
> _______________________________________________
> Lustre-discuss mailing list
> Lustre-discuss at lists.lustre.org
> http://lists.lustre.org/mailman/listinfo/lustre-discuss


Cheers, Andreas
--
Andreas Dilger 
Principal Engineer
Whamcloud, Inc.






More information about the lustre-discuss mailing list