Dear Starwind guys,
I've run into a problem that the second node of a two-node HA cluster does not match the L2 caching policy set up for the storage device created on the first node. The Replication Manager wizard only creates a replica disk image in a location I point it to, and ignores the L2 caching policy I've configured on the first node, not even allowing to set it up manually afterwards.
Therefore the performance of the iSCSI storage device on the secondary node is inferior to that of the primary one, since it has to operate without an L2 SSD-based cache.
How do I set up an HA storage device with the exact same L2 caching policy on both of its two nodes?
Additional details:
- I have two ESXi hosts configured identically: each of them has one HDD, and one SSD datastore
- Each one of the hosts has a StarWind Virtual SAN software running in a Windows Server 2012 R2 virtual machine
- Both virtual machines, in addition to the base OS virtual disk, have two virtual disks attached (one from an HDD-based, and from an SSD-based datastore)
- When creating a standalone non-clustered storage device on the first node, I put the main 100Gb disk image on an HDD-based partition, and a 10Gb L2 cache file on an SSD-based.
- After that I used Replication Manager to turn the newly created device into an HA storage, creating a synchronous two-way replica on the second node.
- To ensure that L2 cache exists for the first node, I checked its SSD drive - and the L2-Storage1.img file was there. However, there was no corresponding cache file on the second node's SSD drive.
The Latest Gartner® Magic Quadrant™Hyperconverged Infrastructure Software
