We use /dev/disk/by-path for this reason, but we confirmed that is stable for our HBAs. Maybe /dev/disk/by-something is consistent with your controller. Cheers, Dan
-- Dan van der Ster || Data & Storage Services || CERN IT Department -- -------- Original Message -------- From: Sidharta Mukerjee <[email protected]> Sent: Wednesday, March 12, 2014 07:51 PM To: [email protected] Subject: [ceph-users] if partition name changes, will ceph get corrupted? If a partition name such as "/dev/sdd" changes to "/dev/sde" and ceph was already mapped to the old "/dev/sdd", how will CEPH react? For example, would it get corrupted, or notice a problem and fail remove that one OSD from cluster or somehow automatically re-adapt? FYI: I ask because we added a new hard disk (disk #3) in the middle of a 8-disk server and I'm not sure if the PERC RAID controller and/or the Centos 6.5 OS will rename the original parititions so that the partitions are names alphabetically or just add a new paritition (/dev/sdh) for this 3rd disk.
_______________________________________________ ceph-users mailing list [email protected] http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
