This post is more than 5 years old
13 Posts
0
1628
high avg, queue depth on a LUN - mismatch between SANHQ and host perfmon counters
Hi. I have a EQL group (3 members) hosting LUNs that are used in Hyper-V clusters as CSVs. One of these LUNs reports (SANHQ) higher than expected avg, disk queue length - the numbers vary from about 6 to 9 steady. However on the Hyper-V hosts that have this LUN connected the current/avg. disk queue length counters are much lower (usually below 1, with occasional spikes to 2 or 3) and nowhere near to what SANHQ is reporting.
I've concurrently ran a live view session in SANHQ for this particular LUN and monitored perfmon counters on the Hyper-V servers and that confirmed me what I wrote above.
What could be the problem? For other LUNs the numbers match between SANHQ and the hosts.
Anonymous
5 Practitioner
5 Practitioner
•
274.2K Posts
0
July 11th, 2017 05:00
Hello,
Anything related around performance issues is very difficult to triage without being able to gather diagnostic info. I would suggest you open a support case. On the LUN affected is there actually a performance issue or just the load shown in SANHQ?
What EQL firmware are you using and what version of SANHQ?
Regards,
Don
Budrumi
13 Posts
0
July 11th, 2017 05:00
Hello Donald,
thanks for the prompt response. It was a mistake on my part as I didn't get a sum info from all the hosts connected to this LUN, when I've done that, the numbers +- match, so it seems it's caused by the load on the hosts, so most probably not an EQL issue and I'll dig deeper on the hosts.
For what it's worth: using SANHQ 3.2 and EQL FW 9.0.5
Anonymous
5 Practitioner
5 Practitioner
•
274.2K Posts
0
July 11th, 2017 06:00
Hello,
Thank you for the update.
FW 9.0.7 and 9.1.1 are now available.
SANHQ 3.3.1 is as well.
Sounds like maybe you have a LUN with much more active VMs, so moving some to another LUN might help better balance things out.
Regards,
Don