Dell R610 (6x disks)
2x Dell 146GB SAS 15k HDDl
4x Pliant LB400M 400GB SSD
You certainly don't need 15K HDDs for a boot pool, but it won't hurt anything other than power consumption.
The Pliant/Sandisk SSDs may end up getting the boot for faster SLOG devices; I'll touch on that later.
Dell PowerVault MD1200 (12x disks)
12x Dell 3TB SAS 7.2k 3.5" 6G Hard Drive 91K8T
Magic decoder ring says "Seagate Constellation ES.2" - the 3TB model was known to be less-than-reliable historically, I'm afraid. Not quite as horrific as the consumer ST3000DM001 (to my knowledge, there was a class-action launched over that drive) but it still wouldn't be something I'd build a backup pool out of. If you don't have any alternative models you can use, make sure to keep several offline spares, set up alert notifications over email, and respond to those alerts promptly. Scheduled scrubs and RAIDZ2 will help mitigate this, but still - tread lightly here.
1x Dell PowerVault MD1220 (24x disks)
24x HPE Intel DC S3520 Series 1.6TB SSD 2.5"
Good choice - the S3520 while still being a "read focused" drive picked up a lot of endurance over the earlier S3500, your drives being rated for just under 3PBW lifetime vs. the 880TBW lifetime of their earlier cousins. The HPE models might hide their total writes in SMART data though - check to see if you can monitor this value.
I have 1x Chelsio 10G NIC card connected to 1 PCI Slot and the second PCI slot has the Dell PERC H200 SAS card in IT mode (screenshot bellow)
How are the internal bays connected? I assume you've still got a card in the "Storage" slot (where the PERC 6i is in the picture) - while a hardware RAID card will be acceptable for a boot device (and possibly even L2ARC) I wouldn't use it for anything like SLOG/meta where data integrity is paramount. You might need to look into the "Moving an H200 into the internal storage slot" process I described in my last post.
Not really keen on having the server down for swapping SLOG
Unfortunately a moot point here as you don't have the PCIe slots for an NVMe SLOG; but is worth noting for the potential R620 upgrade in the future.
If the 4x Pliant LB400M are not good as SLOG due to the speed, will I not be better using then as mirror SLOG on the Spinning disk or metadata? Our main storage is the Dell EqualLogic at this moment so we will be replacing the r610 as sugested but not before Q1 of next year.
If your spinning disks are only handling an asynchronous sequential write workload (such as "store these backups, and hopefully we don't ever need them") then they won't make use of an SLOG at all. Rapid metadata writes are similarly not as crucial in a backup workload - and if you really needed rapid metadata reads, those can be more safely handled by a metadata-only L2ARC device.
With your pool devices having power-loss-protection and fast-ish sync writes, it's
theoretically possible although not necessarily recommended to run sync writes without an SLOG here. I would still recommend that you use a pair of write-intensive SAS SSDs like the WD DC SS530 "WUSTM" model I mentioned above, as the goal is to push write latency as low as possible.
The VMs will be over iSCSI and NFS connection
Will you be choosing "one or the other" or do you have to run both? One of these is a NAS protocol, one is a SAN protocol, and they may have different methods of bundling connections for redundancy and throughput (NFS is okay with LACP - iSCSI isn't)
They also might not take too kindly to sharing the same physical interface in terms of congestion avoidance. If you have to run both at once, I'd suggest tuning the hosts to behave in an active/passive manner, where all of the NFS traffic goes on interface cxgbe0 by default, reverting to cxgbe1 on failover - and iSCSI having the inverse behaviour.
