Take the 2-minute tour ×
Unix & Linux Stack Exchange is a question and answer site for users of Linux, FreeBSD and other Un*x-like operating systems.. It's 100% free, no registration required.

one of my RAM sticks causes a Kernel Panic on my Ubuntu 10.10 (something like "not syncing" with a lot of memory adresses shown on screen). It's definitely this one RAM stick and not its socket because when I put one of the other sticks into the slot of the one RAM stick, everything is ok. How does it come that memtest doesn't find any errors after several cycles but Ubuntu is not able to boot while using this one special RAM stick? Does anybody have an explanation for that?

share|improve this question
3  
You should provide the stack trace or something with more information. I also have problem with 10.10, and the problem also gives "not syncing". I haven't found a real solution, but a workaround is to add acpi=off to the kernel parameter list during boot. –  phunehehe Dec 26 '10 at 10:32
4  
RAM sticks are complex pieces of equipment operating at top speed with no margin for errors, and they are tested for basic defects before shipping. So when you do encounter errors, it tends to be only for certain access patterns. Memtest does its best to try nontrivial access patterns, but any multitasking OS is still more stressful for a marginal RAM stick than memtest and can uncover errors. –  Gilles Dec 26 '10 at 11:29
    
@phunehehe. I can only write the memory addresses down, because after booting with the damaged RAM, the PC is frozen. Is this what you'd like to see? –  Bevor Dec 27 '10 at 17:14
    
@Beavor or take a screencap with a digital camera. –  Josh Jan 13 '11 at 22:58

2 Answers 2

up vote 1 down vote accepted

What is "several" passes? What memtest tests have you run? I know I have seen memtest86+ take up to 6 or 7 passes to find an error with RAM sticks. Also, make sure you run the full battery of tests.

It certainly does sound like the RAM is bad. I too have had not syncing panics because of bad RAM.

share|improve this answer

Are you running memtest with only the 1 (possible) faulty memory module (or a pair if they have to be paired)?

You could probably get a copy of the error report by using the kexec/kdump service, particularly if you can get a copy of the crashdump kernel someplace where the memory error doesn't occur. You could also use the mem=128M kernel parameter to boot a system only using the first 128 megabytes of memory, to see if that gets you a working system.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.