I have questions on SSD pool and ILM mechanism I still did not found answers for after cross reading dozens of web sites and forums.
Probably the response is obvious but I feel like I am missing something here and would like to have certainties on these points.
So, let's suppose I have a 4 hybrid (1065) node cluster and I am running out of SSD.
Q1) Can I technically add an All Flash node in the same cluster container ?
Q2) Will the VMs hosted on this specific AF node automatically be all flash VMs (with no ILM and cold tier drain mechanism) ... or will they use the cluster cold storage pool (from other hybrid nodes) ?
Regarding the ILM mechanism and different SSD thresold, I (think I) understood that:
- data will preferentially be stored on local SSD of each node until it reaches 75% usage
- If one node reaches 75% of SSD usage, then ILM will send "less hotest" local SSD data over the network to other remote SSD nodes
- If all nodes finally reaches 75% SSD usage, then ILM will move (15% of) data to the cold tier
Q3) Regarding this last thresold for draining data to cold tier, is it 75% of the overall cluster SSD pool OR 75% of each of the individual node ?
Q4) Appart from this 75% thresold, is there any other mechanism which can naturally leads to data being invalidating from hot tier ?
Q5) If not, then I guess the SSD usage monitoring in Prism should sooner or latter stabilize around [60%-75%] of global usage for most clusters ?
Is that right or am I misunderstanding something ?
Thanks in advance for your help,
Is AF node also using cluster cold pool & clarifications on ILM thresold
This topic has been closed for comments
(Part of my questions/guess are based around the bible writeup: http://nutanixbible.com/#anchor-storage-tiering-and-prioritization-78)
Already have an account? Login
Login to the community
Login with your account
Enter your username or e-mail address. We'll send you an e-mail with instructions to reset your password.