We added a couple of storage nodes to a modest 3-node cluster that was running out of disk space.
The alerts about running out of storage, and not having enough for redundancy, are gone.
A few months in, and every time I look in Prism, under Hardware, I see that the storage nodes' "Total Disk Usage" has not gone up very much. The compute nodes' disk usage is about 5-6 times as much as the storage nodes (and the compute nodes actually have more storage onboard than the storage nodes).
I understand that Nutanix tries to keep a VM's storage on the same host that provides its compute resources (and the copy is sharded and spread out among other nodes). Is that why I am seeing so little utilisation of the new storage nodes in Prism?
Page 1 / 1
Apologies for late reply. Please let me know if you find the article helpful.
Your statement:
“I understand that Nutanix tries to keep a VM's storage on the same host that provides its compute resources (and the copy is sharded and spread out among other nodes)”
is true in general, and also, “Disk balancing” is triggered if any of the disks go beyond 70% usage, or if the mean usage threshold of the cluster goes beyond 35%.. I believe the storage page in prism should show this.
Hope this help a little.
Regards,
-Said
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.