Hi,
We have a fs with  3OSTS(~200TB)+1MDS,ldiskfs on CentOS6.6 . The upgrading gone smoothly without loosing data. But there is a BUG regarding to logs, I am not able to clean/remove change_logs. The crush of MDT is triggered on lfs clean_logs command.
Our another system with 24OSTs is running with inkernel-2737-gbe60ead-PRISTINE-2.6.32-431.29.2.el6_lustre.gbe60ead.x86_64 since 6 month under heavy load w/o problems.
If you are not using change logs probably it will be ok to upgrade.
With zfs backend I was not able to proceed after the upgrade. Somehow the zfs version 0.6.3-1.2 vs 0.6.3-1.1 are not compatible with the osd_zfs.

One need to be careful with upgrades usually it a dangerous procedure, at least be sure that you have most of the data on the backup.

I am curios if your upgrade will be successful.
cheers,
arman.


***********************************************************
 Dr. Arman Khalatyan              eScience -SuperComputing 
 Leibniz-Institut für Astrophysik Potsdam (AIP)            
 An der Sternwarte 16, 14482 Potsdam, Germany              
***********************************************************

On Tue, Jan 6, 2015 at 10:17 AM, Jérôme BECOT <jerome.becot@inserm.fr> wrote:
Hello,

Can someone here confirm if i'm safe from any risk of data corruption or any further troubles by upgrading my MDS and OSS servers software from the online feature release 2.6.0-RC2--PRISTINE-2.6.32-431.20.3.el6_lustre.x86_64 to the last 2.6.91 version ? or is there some modifications that may lead to a possible crash ?

(proceeding to the usual stop-start services respecting the good order)

We are running test over our new platform here that lasts for more consequent purposes and as the 2.7 will be out there shortly we want to keep it functionnal in 2.6.

Thanks

--
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

_______________________________________________
HPDD-discuss mailing list
HPDD-discuss@lists.01.org
https://lists.01.org/mailman/listinfo/hpdd-discuss