I want to look up, when Storage I/O Control started to throttle IOps on hosts in an ESXi Cluster.
For that I analyse the storageRM logfile and I think I found the correct events, here is an example:
2020-01-08T12:14:24Z hostfqdn.local storageRM: Throttling anomaly VOB for naa.id: 59, 0.203814
Can someone please tell me, for what the red marked values that are named right after the naa id are representing. At first I thought the first value ist the set maximum queue depth (DQLEN), however it sometimes it reaches value that are much lower (example: 3, 0.00112019) than what is shown in the performance metrics in the vsphere client, sometimes the value is much higher than the possible queue depth of 64 on the adapter (example: 168, 0.203217) - so maybe this is the execution throttle\queue depth? (see: http://qgt.qlogic.com/Hidden/support/Current%20Answer%20Attachments/VMware.pdf )
Does SIOC set a larger Queue Depth per LUN or Host, than the defautl maximum of 64?
I try to compare following values:
- the ones from the storageRM log, like here: 2020-01-08T12:14:24Z hostfqdn.local storageRM: Throttling anomaly VOB for naa.id: 59, 0.203814
- esxtop on the selected host: hostfqdn.local > disk device > check DQLEN value for disk device
- vsphere client - select Datastore - Performance - Hosts - Max Queue Depth per Host > check real time value for host hostfqdn.local
The values do not match.
edit: Change topic so that the thread is not moved again