Re: Dell PERC H700/H800 - Mailing list pgsql-performance

From Scott Marlowe
Subject Re: Dell PERC H700/H800
Date
Msg-id dcc563d11002130031g667f3c2s6007d24fba1c3871@mail.gmail.com
Whole thread Raw
In response to Re: Dell PERC H700/H800  (Greg Smith <greg@2ndquadrant.com>)
List pgsql-performance
Recently most of our Dell servers came up for warranty extensions and
we decided against it.  They're all 3+ years old now and replacement
parts are cheaper than the Dell warranty extension.  In fact the price
Dell quoted us on warranty extension was about twice what these
machines are going for on Ebay used right now (i.e. about $800 or for
each server).  I can't imagine that warranty becoming a better value
over time.

Recently, a 73GB 2.5" drive in one of our 1950s died.  The Dell price
was something insane like $350 or something, and they were just out of
warranty.  Put the part number into Ebay and found two guaranteed
pulls for about $70 each.  Ordered both and with shipping it was right
at $150.  So now I've got a replacement and a spare for about half the
cost of the single replacement drive from Dell.  And they both work
just fine.

I now buy hardware from a whitebox vendor who covers my whole system
for 5 years.  In the two years I've used them I've had four drive
failures and they either let me ship it back and then get the
replacement for non-urgent parts, or cross-ship with a CC charge /
refund on priority drives, like for a db server.  Note that we run a
lot of drives and we run them ragged.  This many failures is not
uncommon.

These machines are burnt in, and if something critical fails they just
ship it out and I have the replacement the next day.  And there's no
bumbling idiot (other than myself) turning a screwdriver in my server
without a wrist strap or a clue.

pgsql-performance by date:

Previous
From: Bryce Nesbitt
Date:
Subject: Re: Re: 512,600ms query becomes 7500ms... but why? Postgres 8.3 query planner quirk?
Next
From: Pierre Frédéric Caillaud
Date:
Subject: Re: 512,600ms query becomes 7500ms... but why? Postgres 8.3 query planner quirk?