Update 3/7/2017
This post is a bit dated but still gets lots of traffic. The commands below may or may not still work. This was a bit of esoteric study into the mysteries of an Isilon chassis. Have not worked with Isilon in a while. If anyone has updates please email them to me via the contact form https://realworlducs.com/contact-rwucs/ and I will update and give credit of course.
InsightIQ is a must have for historical and current performance of every kind for an Isilon. Also the main Isilon screen shows a good deal of performance data, and that covers most of what I need daily to keep an eye on things.
Along with all that above like to have a few CLI commands to check status quickly. These are some common commands I run daily. Every workload is different and the number of chassis, clients, w/r balance, and IOPS make a huge difference. Always best to keep an eye on how the Isilon utilization is trending and keep tabs on client performance.
Check load balancing across all nodes in cluster
Sessions per node
isi_for_array “isi smb session list |grep -i computer |wc -l”
Drive utilization for specific chassis
isi statistics drive –nodes=2 –top
Drive utilization for current chassis
isi statistics drive –top
From what I have seen the 3TB and 4TB SATA drives perform fine until about 250-300 IOPS. After that metadata refresh can be slow in directories with over 50k files.
Cluster wide statistics
isi statistics pstat –top
IOPS for Cluster
isi statistics query –nodes=all –stats=node.disk.xfers.rate.sum –top
Disk IOPS per Chassis/Drive
Edit for you drive number per chassis. This is based on 36 drives
for i in {0..35};do isi statistics query –nodes=all –stats=node.disk.xfers.rate.$i;done$i;done
Summary
Check your drive IOPS per chassis. They should all be equal by nature, so figure the average looking at a chassis. Than look at the Disk I/O for cluster under pstat. Figure what the IOPS would be for cluster with drives at 300 IOPS and that is a good measurement for max throughput for cluster using 3-4TB SATA. Note that clusters with less than 5 chassis take a hit and might run high IOPS because OneFS striping has to hit every chassis per write.
Be First to Comment