So I dropped an Intel SSD in our test x4500 last week and have been playing
with it a bit.

Performance wise, it's great. A source code repository that took 18 minutes
to check out into NFS mounted ZFS space only took 3 minutes after adding
the SSD as a slog (the performance was almost as good as simply disabling
the zil). Untar'ing files and other common operations involving lots of
files were also much quicker.

Unfortunately, fma isn't very happy with it :(, it keeps complaining that
the self test failed and marks the drive as faulty. It's not a functional
issue, the drive remains available and works fine, but the chassis fault
light is on, the drive fault light for the bay the SSD is in is on, the
IPMI management drive failure indicator is asserted for that bay, and the
fault management logs are cluttered with spurious false alerts.

I got an official confirmation that the x4540 SSD is basically an Intel
X25-E, probably with different firmware containing the test methods fma is
looking for. I'm not sure what other differences there might be, the price
difference is pretty drastic (I picked up a stock X25-E for about $350,
list price on the x4540 SSD is $1500), but I guess like with the markup on
regular hard drives and memory you're paying for the privilege of having it
under service contract and having full support for any problems that arise.

Even if we could pay that cost (ah, the california budget), since the drive
won't be qualified for our x4500's, there's no way to get to a supported
configuration anyway :(, although presumably the x4540 SSD in an x4500
would at least past self test (anybody tried it?).

On a side note, is there any easy way to get fma to not test a particular
drive?


-- 
Paul B. Henson  |  (909) 979-6361  |  http://www.csupomona.edu/~henson/
Operating Systems and Network Analyst  |  hen...@csupomona.edu
California State Polytechnic University  |  Pomona CA 91768
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss

Reply via email to