I need to test my system's memory and have installed memtest86+. However, I can't find any documentation on basic usage. Specifically, I need to know
(i) How to run memtest86+ and save the list of bad blocks it finds.
(ii) How to mark the bad blocks so that they are not used. (As I understand it, memtest86+ does not do this but there is some other application that can use the bad block list it generates).
I'd appreciate any information on how to handle this and/or on relevant documentation.
Thanks
Scheduled Maintenance: We are aware of an issue with Google, AOL, and Yahoo services as email providers which are blocking new registrations. We are trying to fix the issue and we have several internal and external support tickets in process to resolve the issue. Please see: viewtopic.php?t=158230
[Software] How to use memtest86+ (can't find doc)
Re: [Software] How to use memtest86+ (can't find doc)
Memtest is for RAM: memory. Bad blocks marking so they are not used is for hard disks, not for memory. If memory is faulty, it needs to be replaced.
Re: [Software] How to use memtest86+ (can't find doc)
Thanks, I had the terminology wrong. I've read that there is a kernel option (memmap ?) that can be used to have the kernel skip certain areas of memory. I'm still looking for documentation on it that's current and appropriate for kernel debian uses.
- kent_dorfman766
- Posts: 535
- Joined: 2022-12-16 06:34
- Location: socialist states of america
- Has thanked: 57 times
- Been thanked: 70 times
Re: [Software] How to use memtest86+ (can't find doc)
I've read that there is a kernel option (memmap ?)
That does ring a bell, although I've not had reason to use it.
This is the authority on the subject of kernel boot options:
https://www.kernel.org/doc/html/v4.14/admin-guide/kernel-parameters.html
That does ring a bell, although I've not had reason to use it.
This is the authority on the subject of kernel boot options:
https://www.kernel.org/doc/html/v4.14/admin-guide/kernel-parameters.html
-
- df -h | grep > 20TiB
- Posts: 1401
- Joined: 2012-10-06 05:31
- Location: /dev/chair
- Has thanked: 79 times
- Been thanked: 175 times
Re: [Software] How to use memtest86+ (can't find doc)
You might be able to achieve this by (ab)using memmap parameters to mark certain memory regions as reserved, but that's not really what these features are for so you'll probably have to experiment a bit yourself.
Long ago we had Rick's 'badram' patch for this, but that was back when RAM was so hideously expensive that it justified spending hours testing and mapping regions to exclude... Even then it was never popular enough to make it into mainline.
These days the standard answer really is just "replace the defective module". Unless the memory in question is soldered to some particularly special board, it's unlikely to be worth the time to try and hack around it vs simply swapping out a DIMM.
As for memtest86+, there's nothing stopping you from manually noting the failing addresses it finds, but it isn't going to save a ready-baked file to feed to the kernel - not least because it has no concept of files or disks to save such a thing to. If you really need to capture memtest86+ output directly, IIRC it does support a serial console...
This slice of ancient history may be of interest, and the memtester utility it mentions is, AFAIK, still in the Debian repos. YMMV, but it should do what you're after... At least in theory.
Here is an equally fossilised (though slightly more trustworthy IMO) example of using memtest86+ and hex addresses for the same. From gentoo of course, because, well, gentoo users do weird things.*
There are probably other guides to be found in the archives too, but again I don't think there's a whole lot of interest in this sort of thing these days, so you will at least want to cross-reference with current kernel documentation.
Ed. If you are using GRUB 2.x, there's also this possibility. It's news to me though, so up to you to try it and find out if it still works.
Once is happenstance. Twice is coincidence. Three times is enemy action. Four times is Official GNOME Policy.
Re: [Software] How to use memtest86+ (can't find doc)
Thanks to all. I've never had memory fail before, but I guess there's a first time for everything. Looks like I have to buy some ram.
-
- Global Moderator
- Posts: 3049
- Joined: 2017-09-17 07:12
- Has thanked: 5 times
- Been thanked: 132 times
Re: [Software] How to use memtest86+ (can't find doc)
Be aware that up to bullseye the memtest86+ package works only with BIOS/legacy boot, not UEFI boot.
If the machine boots in UEFI mode, you need the package from bullseye-backports or testing at least (or the upstream version).
If the machine boots in UEFI mode, you need the package from bullseye-backports or testing at least (or the upstream version).
Re: [Software] How to use memtest86+ (can't find doc)
There is an old bug report that tell that badram output was incorrect: https://bugs.debian.org/cgi-bin/bugrepo ... bug=254210
I didn't do deep test related to it (can require big amount of time as need to find faulty ram but only with few sector issue that can be excluded, exclude them and check if work correctly)
I saw on upstream git history that there was some fixes related included in 6.10: https://github.com/memtest86plus/memtest86plus/pull/178
if someone will test it please tell me if sectors reported to set in badram is returned working correctly in 6.10
anyway as someone wrote in major of cases ram need to be replaced FWIK, exclude sectors reported will require more time and it may not be enough to solve the problem. and honestly i never used tried that possibility but I always replaced faulty ram module both at work and outside
I didn't do deep test related to it (can require big amount of time as need to find faulty ram but only with few sector issue that can be excluded, exclude them and check if work correctly)
I saw on upstream git history that there was some fixes related included in 6.10: https://github.com/memtest86plus/memtest86plus/pull/178
if someone will test it please tell me if sectors reported to set in badram is returned working correctly in 6.10
anyway as someone wrote in major of cases ram need to be replaced FWIK, exclude sectors reported will require more time and it may not be enough to solve the problem. and honestly i never used tried that possibility but I always replaced faulty ram module both at work and outside