Re: SSDs, VM SANs & RAID - WAS Re: [gentoo-user] SSD partitioning and migration

[Pandu Poluan]

On Jul 20, 2013 9:27 PM, "Tanstaafl" <tanstaafl@×××××××××××.org> wrote:
>
> On 2013-07-19 3:02 PM, Paul Hartman <paul.hartman+gentoo@×××××.com> wrote:
>>
>> I think you are. Unless you are moving massive terabytes of data
>> across your drive on a constant basis I would not worry about regular
>> everyday write activity being a problem.
>
>
> I have a question regarding the use of SSDs in a VM SAN...
>
> We are considering buying a lower-end SAN (two actually, one for each of
our locations), with lots of 2.5" bays, and using SSDs.
>
> The two questions that come to mind are:
>
> Is this a good use of SSDs? I honestly don't know if the running VMs
would benefit from the faster IO or not (I *think* the answer is a
resounding yes)?
>

Yes, the I/O would be faster, although how significant totally depends on
your workload pattern.

The bottleneck would be the LAN, though. The peak bandwidth of SATA is 6
GB/s = 48 Gbps. You'll need active/active multipathing and/or bonded
interfaces to cater for that firehose.

> Next is RAID...
>
> I've avoided RAID5 (and RAID6) like the plague ever since I almost got
bit really badly by a multiple drive failure... luckily, the RAID5 had just
finished rebuilding successfully after the first drive failed, before the
second drive failed. I can't tell you how many years I aged that day while
it was rebuilding after replacing the second failed drive.
>
> Ever since, I've always used RAID10.
>

Ahh, the Cadillac of RAID arrays :-)

> So... with SSDs, I think another advantage would be much faster rebuilds
after a failed drive? So I could maybe start using RAID6 (would survive two
simultaneous disk failures), and not lose so much available storage (50%
with RAID10)?
>

If you're using ZFS with spinning disks as its vdev 'elements', resilvering
(rebuilding the RAID array) would be somewhat faster because ZFS knows what
needs to be resilvered (i.e., used blocks) and skip over parts that don't
need to be resilvered (i.e., unused blocks).

> Last... while researching this, I ran across a very interesting article
that I'd appreciate hearing opinions on.
>
> "The Benefits of a Flash Only, SAN-less Virtual Architecture":
>
>
http://www.storage-switzerland.com/Articles/Entries/2012/9/20_The_Benefits_of_a_Flash_Only,_SAN-less_Virtual_Architecture.html
>
> or
>
> http://tinyurl.com/khwuspo
>
> Anyway, I look forward to hearing thoughts on this...
>

Interesting...

Another alternative for performance is to buy a bunch of spinning disks
(let's say, 12 of them 'enterprise'-grade disks), join them into a ZFS Pool
of 5 mirrored vdevs (that is, a RAID10 a la ZFS) + 2 spares, then use 4
SSDs to hold the ZFS Cache and Intent Log.

The capital expenditure for the gained capacity should be cheaper, but with
a very acceptable performance.

Rgds,
--