Hello,
As we had a inode usage issue, we did as Andreas advised :
- add a new ost with much more inodes
- disable the first ost
- use lfs_migrate to move files to the new ost
We have about 16M files on this OST (7.5TB). Our OSS are connected via
10Gbps Ethernet, though the lfs_migrate command has been launched on a
1Gbps connected client.
It's running for 5 days and is at about 50% now. It's pretty slow.
Since the command has been launched on a client, does the data transit
through it ? Is there any way to improve performances ? (we will have to
move data from the 2nd ost back once the first one will get more inodes,
then reformat it, then rebalance)
Thank you
--
Jérome BECOT
Administrateur Systèmes et Réseaux
Molécules à visée Thérapeutique par des approches in Silico (MTi)
Univ Paris Diderot, UMRS973 Inserm
Case 013
Bât. Lamarck A, porte 412
35, rue Hélène Brion 75205 Paris Cedex 13
France
Tel : 01 57 27 83 82
_______________________________________________
lustre-discuss mailing list
[email protected]
http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org
