[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Server crashing (revisited)



Hopefully those people who replied to this issue earlier can respond.

If you recall I had a problem with a server crashing (read below). I started doing a memtest on the server itself while operating. It has been running now 3 days. The server has not once crashed that whole time. But this morning I stopped the memtest. Zero errors were reported out of some 300+ tests. Then within minutes, it crashed. Now isn't that peculiar?

Any explanation out there?

Curtis


On 19 juil. 05, at 13:54, Curtis Vaughan wrote:


On 19 juil. 05, at 12:35, TreeBoy wrote:


On Tuesday 19 Jul 2005 19:30, Curtis Vaughan wrote:



The only time this happened to me was when I had broken RAM.

Can you afford to run memtest on the machine for a couple of days
in order to
rule it out.

I swapped the RAM into another machine and it eventually failed
memtest on
the third day of testing - but I haven't had another failure since.
(Fingers
crossed.)

Cheers,



Ok I started running memtest, but I'm not sure exactly how to use it.
First, I tried memtest 100m --log, but that pretty much stopped the
computer from doing anything else. So, I stopped that test and put it
to 50m. But am I doing this right? Oh, by the way I think I have
somewhere around 120 Mb RAM



Hi.

This is why I asked if you could "afford to" do the test.

I was lucky because I had a spare machine and spare RAM: I swapped the RAM and put the "dodgy" RAM into the spare machine. I then used a Knoppix boot CD and
booted into a memtest on the spare machine.

This meant that while the test was happening, I was not otherwise using the
machine.

Sorry.



Yeh, I think I can do that. But if I just leave it testing 50m would it show a problem if one exists? Or do I need to run memtest at full 120m? And finally am I issuing the command right? Or is there some other parameter(s) I should use?


Curtis

PS. The conversation about types of RAM is not bothering me so if it's constructive for others by all means continue.


--
To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.orgwith a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org





Reply to: