HDDs for ZFS (via FreeNAS) - I've a couple questions on the matter

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

XavierMace

Diamond Member
Apr 20, 2013
4,307
450
126
Also remember ESXI doesn't see LUN ID's about a certain number. I don't recall what that is off the top of my head (maybe 2048?), but I know Solaris auto assigned a LUN ID outside the range ESXI recognizes so the host couldn't see the storage. It took me hours to figure that out.

I've had no issues so far with the inception/nesting doll SAN. Just waiting to pick up a few more SSD's before I switch over to it entirely and just use the standalone one for backups. It might not be the simplest solution but it's cheaper than running a second box.

SATA DOM vs USB Thumb Drive: I've had 1 USB boot drive die on me, so I'm running a cheap SATA DOM in the new host. So far so good. Performance wise, not really any faster to boot than the USB drive.
 

thecoolnessrune

Diamond Member
Jun 8, 2005
9,673
583
126
Regarding the NFS comment though... could you explain further? Hasn't ESXi supported NFS datastores for awhile now? Or is there a distinction still between ESXi and ESX in the datacenter?

Just noting here that it's important to remember that IT guys are like any other people. The industry is filled with myths and "This is the way I learned to do it, so it's the best". It doesn't help that it's also an industry perpetuated by constant sales, so you always have your guard up.

That being said, NFS Support came after block support in ESXi. Fast forward to today, and NFS is equivalent, or better than block in ESXi. Pretty much full stop. It's been tested repeatedly since the 5.0 days. There's corner cases where you could use one or the other, especially when trying to hobble along older environments, but New vs. New, there's a lot of reasons to go NFS, and not so many for going block. Infrastructure consolidation, familiar toolset, smaller pane of management, easier to DR, ready to go for hybrid cloud, and the absence of that ridiculous need to run manual Unmap commands in ESXi to reclaim thin provisioned storage.

Hyper-Converged lives on file storage (NFS for ESXi, SMB 3.0 for Hyper-V), with some vendors now allowing block storage to be provisioned for use outside the environment (to allow legacy FC servers for instance to access the Hyper-Converged block storage like a SAN's LUN).

We are in a changing time in the industry, and you do yourself a disservice if you do not work with file-based Storage to Hypervisors in regards to a home lab. Knowing the in's / out's of block is extremely important right now too, because there's lots of environments still using it. But File-based storage is the future, and every tech refresh through companies will increase it's share in the market. Definitely don't skip on it.
 
  • Like
Reactions: destrekor