Details
-
Bug
-
Resolution: Duplicate
-
Major
-
None
-
7.3
-
None
-
We are taking snapshots of running vm every night on a pool of 3 Xenservers 7.2 (fully patched) connected to 2 iscsi SR with multipath.
Description
Nighly snapshots of running vm are randomly corrupting iSCSI SR volume groups. When it happens, we need to restore volume group manually with 'vgcfgrestore'.
It already happened 2 times this month :
Mar 14 22:19:56 xenserver08 SM: [31876] ***** LVHD over iSCSI: EXCEPTION <class 'SR.SROSError'>, The VDI is not available [opterr=Command ['/sbin/lvs', '--noheadings', '--units', 'b', '-o', '+lv_tags', '/dev/VG_XenStorage-6010cef0-b5ef-a604-bfd3-a1fde94d0d6f'] failed (5): /dev/disk/by-scsid/23237343363613338/mapper : Checksum error
Mar 20 23:24:11 xenserver04 SM: [32119] ***** generic exception: vdi_delete: EXCEPTION <class 'SR.SROSError'>, Failed to load VDI [opterr=rvgstats failed error is 5]
We have a second pool of XS 7.2 running exactly the same backup process on the same iscsi target model every night for one year now without any problem. This pool has only 2 nodes, I wonder if the problem could happen only on 3 nodes pool.
Best regards,
Nicolas Michaux