freenas ctl_datamove aborted
Description
Problem/Justification
Impact
SmartDraw Connector
Katalon Manual Tests (BETA)
Activity
Dmitriy November 11, 2019 at 8:39 PM
OK I understood. Thank. I will try on the host that makes backups on this Lun to reduce the depth of the queue. Esxi iscsi advanced settings.
RaidZ is not fast (ISCSI NFS). L2ARC does not help. DISKs for backup 2.5" 5TB SATA 5400k
Alexander Motin November 11, 2019 at 6:29 PM
Whether there is a danger depends how initiator handle those errors. It is definitely not good in general.
Our main recommendation is not use RAIDZ for block storage (iSCSI), preferring mirrors. You may also try to limit request queue depth on initiator side.
Dmitriy November 11, 2019 at 6:11 PMEdited
yes, i have 2 raidz of 3hdds. except that it is not fast is there a danger? what are the recommendations? sas disk? on vmware in the iscsi settings there is a delayedAsk setting. It's on. If I off delayedAsk, will it help? or maybe use raidz+raidz ?
Alexander Motin November 11, 2019 at 3:32 PM
It usually means the respective pool is not fast enough for the workload, that makes initiator to abort requests on timeout. As I see you have RAIDZ of 3 HDDs there, that is not the fastest configuration from IOPS perspective.
Dmitriy November 11, 2019 at 2:20 AM
debug attached
what is this ? how fix this?
FreeNAS-11.2-U6 + 2nic 10G in Lagg + vmware iscsi