How do you check the health of the memory?

[Nocriton]

Hi everybody

A few weeks ago, we've got a esxi server freezing.

The system has 144 GB of Ram - and after the second freeze the server noticed that a memory dimm failed.

We've now got the discussion that we should check the health of the system before we exit the maintenance of the system.

This time we left the host for 7 days in maintenance, on the 8. day the system crashed again...

I tired to find some simple piece of bootable software to check the whole 144 gb of memory.

But the memcheck i found is limited to 64 GB of ram.

What kind of software are you using to check the health of your memory?

Best regards

Nocri

[vmroyale]

Hello.

What kind of server is this?  Are there any vendor utilities that can be used?

Good Luck!

[DSTAVERT]

Did you try the one from http://memtest.org

[Nocriton]

Well, our hardware vendor was checking the hardware after the first freeze.

But they we're not actively checking the memory...

But it might be a good option to ask them if there are additional tools avaible.

I already looked at memcheck86 and memcheck86+, but both seems to be limited to >64 GB of memory.

[NTShad0w]

Nocriton,

for a test a large number of memory You should use Memtestx86 version 4.20, it's can be downloaded from:

http://memtest.org/download/4.20/memtest86+-4.20.iso.zip

I've 96GB own custom server and it's working perfectly, soon I'll have 144GB too so eventually I'll do a test again and give You an info, but with 96GB its working.

regards

Dawid Fusek

[shishir08]

To find out (before you buy) which memory is supported by your computer,  you could go to a memory seller like crucial.com or kahlon.com and use  their "advisor tool" or "configurator".

http://www.crucial.com/

http://www.kahlon.com/memory.asp

Plug in your computer information and they will tell you what kind of  memory you need, if modules can be mixed or if they have to be in  matched pairs, whether your system is backwards compatible, etc.  Once  you have this info you can shop around for best prices.

If you want to test just your memory, this free tool is probably one of the most recommended and highly respected available:

http://www.memtest.org/

And another free diagnostic tool for your entire computer is:

http://belarc.com/free_download.html

[Nocriton]

Hi

We found out that the server hat some internal memory testing tool, which is able to test up to 1 TB memory.

So we used this tool to test the memory...

Thx

[SimonStrutt]

Probably worth noting that some vendors own memory checking software tends to be a bit brief.  I've known them to state that everything is fine (despite an earlier PSOD suggesting a prob) and then found errors with a more thorough check.

Memtest is very good, it will take a long time (24hrs+ if you've 100GB+ RAM), but it leaves no stone unturned!

[DoDo201110141]

In my experience, memtest also spins up the CPU, at least one core.

Therefore gives the CPU a good stability test too. Generally not an issue on server class machines though

[NTShad0w]

Mates,

I can confirm MEMTEST v4.20 working very well on 144GB of ram/memory. HW is 2x intel L5638 on Supermicro X8DAH+-F.

its a slow witch such mem byt very well tool, thanks for an author of it ;-))

regards

NTShad0w