Hi,
I understand a 2 node cluster is effectively network RAID 1 across the two nodes, but what about on each of the nodes?
Do I still use a HBA for the vSAN capacity drives?
What RAID level does vSAN apply to each of the individual nodes?
What are the capacity drive limits per node when in a 2 node cluster?
Hello therealhostman,
"So to clarify your response, when a VM is running on a node that is part of a 2 node cluster, the storage redundancy on that active node is RAID0, regardless of how many capacity drives I have in the node?"
No, it is not RAID0 in the way you suggest, this is not a traditional active/passive setup - it only reads locally, writes to both data-replicas and if/when a capacity-tier device fails it switches reads and writes to the data replica on the other site (~5-7 seconds stun and failover time), it then proceeds to rebuild the data lost from the original site from the existing replica (provided there is adequate slack-space available which I strongly advise doing).
"How does one estimate the usable space of a 2 node cluster, with identical spec nodes where I have say 1 cache device and 5 capacity drives per node?"
As Joerg said, it is going to be RAID1 which is 2x the used VM space (as the Guest-OS sees it) mirrored on each storage-node - there is no RAID5 option in a 1+1+1 or 2+1 cluster as this requires a minimum of 4 Fault-Domains/nodes.
Bob
Hello therealhostman,
"I understand a 2 node cluster is effectively network RAID 1 across the two nodes, but what about on each of the nodes?"
No, you cannot local-RAID across Disks/Disk-Groups on a single node - local protection requires a minimum of 3 nodes on each site/Fault-Domain.
"Do I still use a HBA for the vSAN capacity drives?"
Yes of course - they still have Disk-Groups just like any other vSAN cluster.
"What RAID level does vSAN apply to each of the individual nodes?"
Data Objects are stored as RAID1 across the sites/nodes and then striped RAID0 on each site if required (e.g. size of Objects >255GB or Stripe-Width in Storage Policy= >1).
"What are the capacity drive limits per node when in a 2 node cluster?"
They are the same as any other vSAN node - max 5 Disk-Groups each with max 7 capacity-tier devices per node.
Bob
So to clarify your response, when a VM is running on a node that is part of a 2 node cluster, the storage redundancy on that active node is RAID0, regardless of how many capacity drives I have in the node?
Writes are synchronously written to the 2nd passive node, but the active node storage redundancy troubles me somewhat.
How does one estimate the usable space of a 2 node cluster, with identical spec nodes where I have say 1 cache device and 5 capacity drives per node?
Yes, its a RAID0 for the single member of a 2 node Cluster.
Do determine the usable storage you only take the capacity drives into account. The Cache/Buffer device is not in the game here.
With an FTT=1 its a RAID1 so you get 50% of all the capacity drives.
Regards,
Joerg
Hello therealhostman,
"So to clarify your response, when a VM is running on a node that is part of a 2 node cluster, the storage redundancy on that active node is RAID0, regardless of how many capacity drives I have in the node?"
No, it is not RAID0 in the way you suggest, this is not a traditional active/passive setup - it only reads locally, writes to both data-replicas and if/when a capacity-tier device fails it switches reads and writes to the data replica on the other site (~5-7 seconds stun and failover time), it then proceeds to rebuild the data lost from the original site from the existing replica (provided there is adequate slack-space available which I strongly advise doing).
"How does one estimate the usable space of a 2 node cluster, with identical spec nodes where I have say 1 cache device and 5 capacity drives per node?"
As Joerg said, it is going to be RAID1 which is 2x the used VM space (as the Guest-OS sees it) mirrored on each storage-node - there is no RAID5 option in a 1+1+1 or 2+1 cluster as this requires a minimum of 4 Fault-Domains/nodes.
Bob
Thanks, I understand this better now.