Rockstor with HP Smart Array P410 in RAID5

dolphs · October 21, 2015, 10:14am

Hi - New to Rockstor and willing to give this a go when I am confident following hardware is supported.

Initially I checked HP website and it seems RHEL6 is only supported for their ACU ( HP Array Configuration Utility ) as well ADU ( Diagnostics ), as I like to build my personal cloud with an " HP Smart Array P410 " it is mandatory to have access to these utils, yet I believe this should not be an issue on CentOS7, thus also no issue for Rockstor hopefully installing the RPMs. Once this is confirmed I need to have an answer on following questions …

Rgd the RAID controller I like to attach 8 Seagate Archive HDD v2 ST8000AS0002, 8TB disks configured in RAID5 so 56000 GiB will be available to my cloud, having in theory 7x read speed and no write speed gain. As these drives use Shingled Magnetic Recording (SMR) it might become tricky once rebuilding the RAID-array, but for home purposes I guess the rebuild may take up a night or even a weekend … The cloud server is not meant to be up 24/7 and only needed once access is needed, so WoL is needed as well automatic shut down after eg 20-30 minutes …
As Rockstor uses the " revolutionary " btrfs file system I wonder if one has experience with these drives in RAID5 mode ( with a hardware controller )…

Rest of hardware I like to reuse, assuming the PSU should be able to do its job fine ( estimated around 200W ):

CPU: i7-3770t ( Ivy Bridge )
Motherboard: ASUS P8-h77i
Memory: 16Gb Kingston ( non ECC as board does not support ECC ) 1600MHz
HP P410 Smart Array
LC-Power silver class PSU 300w

The boot disk will be a 500Gb 7200rpm drive ( s-ata600 ).

Wondering if this set up is indeed valid
thanks!

roweryan · October 21, 2015, 11:23am

Hi.

btrfs doesn’t use a raid controller, so to use btrfs in a normal setup, you would bypass the raid controller or leave the raid controller doing nothing…

phillxnet · October 21, 2015, 12:47pm

@dolphs Welcome to the Rockstor community. My main concern would be btrfs’s support for the SMR drivers:-
ie from:- http://www.spinics.net/lists/linux-btrfs/msg48072.html

"I assume you have the host-managed SMR drives. This type needs tweaks to
the operating system so the write patterns play well with the SMR
constraints. Btrfs does not support that out of the box, but my
colleague Hannes Reinecke managed to get it working with some minor
changes to the allocator and disabled writing of superblock copies.

For full support of SMR we’d have to change more than that, currently
nothing prevents to write “backwards” in a given chunk that is allowed
to be written only in the append way. So you can get mixed results when
trying to use the SMR devices but I’d say it will mostly not work.

But, btrfs has all the fundamental features in place, we’d have to make
adjustments to follow the SMR constraints: …"*

and he (KDave a btrfs developer) goes on to link to his notes on the matter:-

github.com

kdave/drafts/blob/master/btrfs/smr-mode.txt

SMR mode
--------

add support for SMR devices, respecing the constraints

constraints:
- only a portion of the disk can be updated randomly
- most of the device is appendonly per chunk
- chunk size is variable, can be 256MB or 1G
- the chunks are aligned
- the requirements are strict and lead to error if not adhered to

other constraints:
- the filesystem must be readable anywhere, it's not guaranteed to be writable
  again if modified outside of SMR environment
- it would be desirable to shape up any filesystem to be SMR-compatible
  (userspace relocation, live relocation with the smr mode enabled, reloc
  should be automatic)

proposed changes:

This file has been truncated. show original

I saw this SMR question come up a little while ago on the linux-btrfs mailing list so just searched for it and pasted here. I do not know myself if this is relevant to your drives ie the “host-managed SMR drives.” bit. The original poster, Warren Hughes, states:-

"> Hi guys, just added a new Seagate Archive 8TB drive to my BTRFS volume
> and I'm getting a tonne of errors when balancing or scrubbing."

I would suggest that you read the linked post and the entire thread if possible as I’m afraid these SMR drives are an unknown property to me.

Oh and what @roweryan said. ie JBOD mode.

We have another forum thread that has some discussion on the Smart Array controllers and a quick look at the links provided there (example) it appears that your HP410 is not a cciss (depricated as of RHEL 7) only device so that’s good news. So at a guess I would say you are fine with that controller as it should default to the build in and upstream supported hpsa driver.

Hope that helps.

dolphs · October 21, 2015, 2:59pm

both of you thanks for your response - In other words, support for the hard drives is on its way, but not yet. Therefore I will remain patient and read the threads. OK thus rest of hardware would be fine ( as I was hoping for ) incl HP RAID controller, thanks again!

felixbrucker · October 21, 2015, 9:32pm

i am using the same drives in my zfs array (raid6), they will work ok for archiving (and just that, store once and never change) with software raid. Problems appear when using hardware raid as the disks tend to take their time when shingeling and get removed from the raid. with almost any software raid these times beeing in the reasonable range for the disks they dont get removed as faulty. If you plan to use them for anything else than archiving they will have a very short lifetime.

HittingSmoke · October 21, 2015, 10:36pm

The P410 does not support JBOD. That means you’d be forced to create the RAID array on the controller, negating all of the RAID management features of BTRFS, or you’ll have to set up multiple RAID 0 arrays to put into a software RAID5 array with BTRFS. Neither situation is ideal.

I have a similar SmartArray model in my server and it’s very frustrating. You do not want to use this controller for Rockstor.

A better solution would be to get a cheap fake RAID controller that will present the unmolested block devices to the OS. Otherwise you’re missing out on the entire point of running BTRFS RAID.