We have had this from 3rd party memory suppliers (on Linux & Windows) - the problem seems to be that the memory isn't properly burned-in or soak-tested. The chips are in spec when they go out the door, but a few hours run-time, and they have to be downgraded. I suggest you try running them at 100MHz, then 66MHz (and even on a 33MHz bus, if you can) - just to test to see if they are out of spec or completely failing. The rule for memory, I'm afraid, is "low cost, low quality". (The generic memory supplied by Insight (our main supplier) is fine.) Regards, Paul. ----------
From: npauli <npauli@st-johns.org.uk> To: suse-linux-uk-schools@suse.com Subject: [suse-linux-uk-schools] More than 64Mb memory and linux Date: 17 June 2001 20:16
Dear All,
Late last week I installed an extra 128Mb memory on the linux box I use to run squid to take it up to 192Mb.
Disaster!!!
Over the weekend the server crashed. I couldn't get it to reboot sensibly. When I did get it going something was steadily grabbing more and more memory and I couldn't even start up a terminal to look for the process.
I tried to reinstall Mandrake 7.2 and then Redhat 6.2 and with both the installs just locked up or crashed.
So I've taken the new memory out and am reinstalling Mandrake 7.2 and this time it seems to be working.
Does anyone know where the problem might be? On each install the all the memory was detected. The memory was of the right sort but was fairly cheap stuff (major on 3rd - whatever that means) from www.redstore.com . For a while the added memory worked well and made an appreciable difference to the machine - but now...
Should I upgrade to the 2.4 kernel?
Thanks for any advice anyone can offer on this.
Nigel Pauli St. John's School, Northwood
-- To unsubscribe, e-mail: suse-linux-uk-schools-unsubscribe@suse.com For additional commands, e-mail: suse-linux-uk-schools-help@suse.com