Bug?: Minor numbers in true evms volumes are *not* persitant!

Rainer Krienke <[email protected]>
Newsgroups gmane.linux.evms.devel
Message-ID <[email protected]>
Hello,

I am running two files servers with SuSE SLES10 (kernel 2.6.16.27, evms 
2.5.5). The (NFS) file servers have disks (5TB hardware raids) attached and 
on each raid I manage the storage available using evmsgui. I create EVMS 
volumes (true EVMS vols, based on the LVM2 region manager). This basically 
works fine.

However lately I had to reboot the file servers and after this reboot some NFS 
clients which had not been rebooted showed a strange behavior. They tried to 
NFS mount (or still had mounted) a volume say server1:/export/evms_volume1 
to /import/volume1 from the rebooted file server. The mount was ok but the 
strange thing was that the mounted volume on the NFS client did not contain 
what server1:/export/evms_volume1 contains but instead the contents of say 
server1:/export/volume3. So the mount was sucessful but the contens was from 
another NFS volume exported by the same fileserver.  This really turns out to 
be a nightmare.

I tracked this bug down to the fact that after a reboot "sometimes" the minor 
device numbers of the NFS exported EVMS volumes in /dev/evms/ simply change. 
Since the major/minor numbers are used in NFS to identify a NFS volume, NFS 
clients suddenly mount the wrong one.

Today I set up a test system and the effect was there again, eg after I had 
created some EVMS volumes based on the LVM2 region manager using evmsgui and 
then rebooted the machine. One or two of the 5 EVMS volumes I previously had 
created suddenly had *different* minor device numbers than they had before 
the reboot.

Is this desired behaviour in EVMS? I cannot imagine but since it happens on 
several machines I really ask myself. If its not desired does anyone have an 
idea what might happen or how the minor numbers of existing EVMS volumes are 
assigned and when they might be reassigned? This might help me to find out 
whats wrong. On the machine at boot time first evms_activate is started by 
the boot script boot.evms then later boot.lvm is started. So evms scannes all 
disks and then a little later lvm does the same again. Could this be the 
cause of this strange trouble?

The bug seems to occur whenever an EVMS volume has been created/deleted and 
the machine is then rebooted. 

Any ideas, or hints how to go on searching?

Thanks
Rainer Krienke

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

_______________________________________________
Evms-devel mailing list
[email protected]
To subscribe/unsubscribe, please visit:
https://lists.sourceforge.net/lists/listinfo/evms-devel
signature.asc (application/pgp-signature, 189 B)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)

iD8DBQBFmqOialdtjc/KDEoRAoXqAJ4uZjaDfU/hKRhpDEVsFUckwQSqgACfULB+
DdayrwsHEaIr6vyhhhhdKaA=
=pNfV
-----END PGP SIGNATURE-----
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.