VM corrupting on iscsi multipath storage me4084

Hi people,
I tested it with version 5.4-3 it is working normally with the discard set in the vm, when it cleans the data in the vm it is automatically cleaned in the storage no longer corrupting the vms.

However on hosts pve I sometimes get the messganes below. Should I worry about it?

Jul 1 21:14:02 pve01 kernel: [21784.699165] blk_cloned_rq_check_limits: over max size limit.
Jul 1 21:14:02 pve01 kernel: [21784.707611] blk_cloned_rq_check_limits: over max size limit.
Jul 2 00:31:01 pve01 kernel: [33604.276164] blk_cloned_rq_check_limits: over max size limit.
Jul 2 03:48:44 pve01 kernel: [45466.632436] blk_cloned_rq_check_limits: over max size limit.
Jul 2 03:48:44 pve01 kernel: [45466.642556] blk_cloned_rq_check_limits: over max size limit.
Jul 2 07:55:24 pve01 kernel: [60267.099581] blk_cloned_rq_check_limits: over max size limit.
Jul 2 08:11:44 pve01 kernel: [61247.378333] blk_cloned_rq_check_limits: over max size limit.
 
Hi,

have you ever found a solution to the issue? We are having a similar issue using a DELL Storage, ISCSI LUNs mounted on 10 hosts (~ 120 running vms). We haven't been able to pragmatically replicate the corruption, but our CI/CD servers filesystems (ext4) are crashing every few weeks. We had a few crash on other VMs but we havn't found any correlation. It seems like the must impacted VMs are the one with the highest IO.

We are trying disabling fstrim on one of CI/CD server and check the next crash... We don't really know where we could check.
 
You must log in or register to reply here.
Top Bottom
Back