Date: Sat, 27 Jul 2002 15:22:17 -0600 (CST) From: Ryan Thompson <ryan@sasknow.com> To: George Ramos <george@durham.net> Cc: freebsd-isp@freebsd.org Subject: RE: Hard drives keep dying.....loooking for some advice Message-ID: <20020727143804.O61380-100000@ren.sasknow.com> In-Reply-To: <005501c235ad$7f954ae0$20fea8c0@ddd>
next in thread | previous in thread | raw e-mail | index | archive | help
George Ramos wrote to 'Ryan Thompson':
> WOW Ryan, you are a nice guy!! You really went out of your way to
> explain things, I REALLY appreciate it.
Not a problem. I help where I can. :-)
> Let me answer some of your welcomed comments.
>
> I agree with you that there is more in play than temperature here.
> Exactly what are you referring to when you say: "Check voltages.
> Regulate power."?
A common cause of apparent failures is irregular power. If your
equipment is not getting consistent power within a few volts of spec,
you might see these sorts of problems. You should have your equipment
connected to a line-interactive UPS (battery backup that also flattens
out power peaks and dips to ensure continuous clean power). You should
also check the voltages received by your motherboard to ensure that
they meet the +/-{5,12}V specs, to rule out flaky power supplies.
Re-check connections/seatings inside the case as well.
> What should I be doing?
>
> We have not only replaced components, we have SWAPPED ENTIRE
> servers, so I don't think that the problem is related to other
> components, we are actually talking about two servers here, both
> with the same effect......
Ok. Make sure you also try different hardware. You might be running
into subtle compatibility issues with the components you are using on
multiple systems. Many a time I have seen people say "I've tried
*everything*", but are still using two boards with the same flaky
chipset on both machines, for instance. In other words, don't use
another identical {controller|board|RAM|...}, try it on a completely
different hardware configuration. Even on crappy hardware, your
application might be slow as hell, but it shouldn't crash stable
hardware.
> As far as the applications go, the problem started when we put on
> line our new virus scanner application (Amavis/McAffee). I guess it
> is possible that my problem could be due to all the "new" extra work
> that the server is doing......
If there is indeed a definite corelation there, then you'll really
need to track down what the software is doing differently with your
system. That being said, software can't destroy hardware, but it *can*
excarbate already-failing hardware if it produces excessive load.
Without knowing what your developing skills are, I would just suggest
monitoring your run queue, open file descriptors, reads/writes per
second, etc. Check logs. Record crash output (if the system ever
panics).
> I wish I could spread the load but we are running in a very tight
> space and I can't add anymore servers where I'm at. Furthermore,
> adding more equipment would worsen the temperature issue......
Hmm... For temperature, determine if things really *are* heating up.
Pop the cover while it's running under load, and see if anything is
too hot to touch. Just be careful doing this.
> The only other servers that I have in there are a gatekeeper and a
> Frontpage server (a windows PC). I guess we could make the Frontpage
> server a BSD server, move a couple of the services from the main
> server to it and make Frontpage run under BSD (that way I would be
> able to move a couple of services over.....) What do you think?
> (We've never set up MS-Frontpage on a BSD box but I think it's
> doable....)
Definitely doable.
How much mail (size, as well as number of messages) are you
processing? You'd have to be scanning a *lot* of messages to make that
much of a difference (multiple messages per second), I would think.
Hope this helps,
- Ryan
--
Ryan Thompson <ryan@sasknow.com>
SaskNow Technologies - http://www.sasknow.com
901 1st Avenue North - Saskatoon, SK - S7K 1Y4
Tel: 306-664-3600 Fax: 306-244-7037 Saskatoon
Toll-Free: 877-727-5669 (877-SASKNOW) North America
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-isp" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020727143804.O61380-100000>