#Moved NFS volume resulted in massive amounts of Other IOPS

1 messages · Page 1 of 1 (latest)

feral plaza
#

Doing some LCM work and moved an NFS volume (vSphere datastore) to new aggregate and node and after the move I'm seeing huge (400k) amount of Other IOPS with low (as expected) amount of read/write.
I see the load in Grafana volume/node dashboard and statistics volume show, but not in statistics top client show so I can't point to a specific client.

Anyone have any idea what is causing the Other IOPS?

I tried moving the volume to another new aggregate and the IOPS stopped when the vol move was done but came back again after two hours.

feral plaza
#

Tried looking for what kind of Other IOPS it is with this line but it is showing nothing abnormal:
statistics start -sample-id <samplename> -object volume -instance <volname> -counter nfs_read_ops|nfs_write_ops|nfs_other_ops|nfs_access_ops|nfs_getattr_ops|nfs_lookup_ops|nfs_total_ops

lyric moat
#

Is the volume deswizzling?

set d -c of;run * wafl scan status;set adm -c on
feral plaza
brisk sage
#

We have seen this for years on multiple volumes (NFS and CIFS) and have never been able to correlate it to anything. Multiple support cases have pointed it to something running on the clients, but we have never been able to identify any virus scans or anything like that that are running. I eventually just gave up and just accept it as the norm, but it really confuses things when you're looking at average/max IOPs and try to plan around it.

feral plaza
#

Damn, that makes it a lot harder yeah.
I don't think my issue is protocol related but I have registered a support ticket and will echo the findings to here.