PowerEdge R720xd w/PERC H710

[zetabax]

Hi there,

I'm new to TrueNAS / FreeNAS and hoping someone can please explain why HBA / passthrough / IT mode is so important TrueNas /FreeNAS?

Dell's PERC H710 raid controller (like their other raid controllers) is a rock solid hardware raid controller so why do I want to enable passthrough which ultimately shifts the overhead onto the CPU?

Thanks

 

[WB3FFV]

I remember reading a good article about this once, but in short ZFS handles the drives itself, and can heal and deal with failures even better than a RAID controller ever could. If your RAID card ever messes up and corrupts anything, ZFS won't know about it, and your array is toast! If I can find the article I will send you a link.

Just FYI, I just finished building up a R720xd LFF server with TrueNAS 12 and it's working great. It's pretty easy to load in the IT software, but if it's something you don't want to do, you can pick them up already programmed on eBay dirt cheap..

 

[jgreco]

@WB3FFV 's answer is correct as a meta-answer. ZFS *is* your RAID controller, and, yes, it uses the host CPU to do that. But it will happily create you an array of hundreds of disks as a single high performance pool, something that your PERC cannot do.

The comprehensive answer to why you won't be using the H710 is available on the forum already.

https://www.truenas.com/community/t...s-and-why-cant-i-use-a-raid-controller.81931/

Also, speaking as someone who's worked with a bunch of PERC's, Dell's PERC H710 blows chunks. Just within the last month, admittedly this on H700's not H710's, I discovered the hard way that Dell has disabled the ability to replace an HDD with an SSD, a feature which the underlying silicon supports. So this turned what should have been an hourlong site visit to a data center into a six hour catastrophe where instead of just simply replacing a HDD in each RAID1 datastore with an SSD, letting it rebuild, and then replacing the other HDD with the other SSD, I had to migrate all the VM's off a hypervisor, shut it down, destroy the entire RAID configuration (three RAID1 mirrors in each), install SSD's, manually configure the new SSD RAID1's, and then reload the hypervisors, one at a time, for each of eight hypervisors. I cannot thank the jerks at Dell enough for making what should have been a trivial upgrade into a late night sleep-robbing ordeal.

I also vote "perccli blows chunks."

 

[zetabax]

Very helpful. Thanks for taking the time to write back.

 

[artlessknave]

I cannot thank the jerks at Dell

artificial limits are such BS.