Live Memory Tester for UNRAID

A diagnostic plugin that lets you test your RAM without needing to reboot.

 

Identify memory problems in situations where a restart is not (immediately) possible or just stress test your RAM.
This plugin provides a convenient GUI for in‐place memory testing using the veteran command-line utility memtester.

 

Installable via Community Applications

 

Edited June 14, 20242 yr by Rysz

Cool ui addition idea, i can see this being useful when people have system instability issues.

did a run on 5 gb and seems to work as expected 👍

Hi,

 

tested this. Works good. Tooks about 4 Minutes for 1 GB. It´s using only 1 Core of the CPU with a 100%.

 

[14.06.2024 19:35:32 CEST] /usr/bin/memtester 1G 1 ============================================================================================= The detailed error output was enabled, please watch this panel for any occuring errors. ============================================================================================= [14.06.2024 19:39:10 CEST] The operation has finished without errors.

[14.06.2024 19:39:10 CEST] Code: 0 - OK.

 

It would be interesting to see how this plugin behaves on a system that is known to have RAM issues.

2 hours ago, UNRA1DUser said:

Hi,

 

tested this. Works good. Tooks about 4 Minutes for 1 GB. It´s using only 1 Core of the CPU with a 100%.

 

[14.06.2024 19:35:32 CEST] /usr/bin/memtester 1G 1 ============================================================================================= The detailed error output was enabled, please watch this panel for any occurring errors. ============================================================================================= [14.06.2024 19:39:10 CEST] The operation has finished without errors.

[14.06.2024 19:39:10 CEST] Code: 0 - OK.

 

 

Thanks for reporting back, it using only one core is normal as the memtester program is not multi-threaded. 🙂 

 

Edited June 14, 20242 yr by Rysz

Hello @Rysz I start testing 4 days ago for 12GB. I have 2x32GB on my system. How log that should test? Can I make plugin upgrade while it's testing?



Regards.

Edited June 17, 20242 yr by MarianKoniuszko

You didn't set a number of loops so it will run endlessly, did 100 full runs so far with no issues... i.e. you're good, just stop it. 

Edited June 17, 20242 yr by Kilrah

17 minutes ago, Kilrah said:

You didn't set a number of loops so it will run endlessly, did 100 full runs so far with no issues... i.e. you're good, just stop it. 

Thanks, so 1 loop is enough?

21 minutes ago, MarianKoniuszko said:

Thanks, so 1 loop is enough?

 

Usually people let memory tests run for around a day or two at most, with 100 loops you'll be good.

I am curious how it works.

To test all of the memory doesnt it have to be made available to test?

How can it do this while unraid is running?

13 minutes ago, _cr8tor_ said:

I am curious how it works.

To test all of the memory doesnt it have to be made available to test?

How can it do this while unraid is running?

 

It can't test all of the memory, but it can test a large subset of your memory (e.g. 12GB out of available 16GB).

It's very useful for situations where you have no physical access or just want to do a quick preliminary test over a suspicion.

 

As an example, today we suspected RAM problems for a Discord user with OS instabilities and filesystem corruption.

Using this plugin, we quickly identified that their RAM had severe problems, becoming apparent within seconds of testing:

 

So it's not a full replacement for an extensive offline memory-test (e.g. using memtest86), but it can be helpful in situations. 🙂 

 

10 minutes ago, Rysz said:

So it's not a full replacement for an extensive offline memory-test (e.g. using memtest86), but it can be helpful in situations. 🙂 

 

Understood, about what i thought but was curious. Thanks for sharing!

1 minute ago, _cr8tor_ said:

Understood, about what i thought but was curious. Thanks for sharing!

If you want to test “all” you should probably just choose memtest on boot screen. I’m not sure that really tests all since at least a small portion is probably used for memtest but it should come close.

Nice utility! I just tested a system of mine which has exhibited corrupt cache drives over the course of the last year and i've just been limping it along. Failures right at the start of testing! Thanks for making this plugin so easy to use.

So my server has been randomly going unresponsive, needing hard reset, since a rebuild

I guess this shows the issue, thank you:
 

  

11 hours ago, psychological-upland7525 said:

I guess this shows the issue, thank you:

 

 

Seems to. Try a real barebones memtest from the boot menu, assuming you have multiple ram sticks try one, if it fails only another etc unless you can isolate a bad one.

Edited October 25, 20241 yr by Kilrah

Thanks for creating this!

 

I wonder if it's possible to extend this to help guard against non-ecc memory silently having a fault which is not obvious and causes a heap of corruption, eg for cheap homelab servers that are not ECC capable.

It'd be awesome if we could use this  schedule this to run once a week or month, do a cycle or two, and have it generate a notification and unraid dashboard error if there's an issue - is that something you'd be open to expanding it with?

 

Alternatively perhaps having the option in the UI to schedule a reboot and execute a 2-cycle offline memtest, writing the output somewhere that can be actioned once the server comes online?

Great plugin, thank you!

On 12/25/2024 at 5:56 AM, maelstrm said:

It'd be awesome if we could use this  schedule this to run once a week or month, do a cycle or two, and have it generate a notification and unraid dashboard error if there's an issue - is that something you'd be open to expanding it with?

Why do you need it? Ram has life warranty now, it should be tested once and forget until you decided to sell it.

3 hours ago, d3m3zs said:

Great plugin, thank you!

Why do you need it? Ram has life warranty now, it should be tested once and forget until you decided to sell it.

Because like all parts RAM can fail at some point in its life. And unlike most PC components, a failure of RAM won't just case your server/PC to not boot, it can cause it to start silently corrupt your data.

 

I booted up a server recently, O/S started fine, seemed OK, but I decided on the off chance to run a RAM test and found that one of the RAM sticks had failed. If I had not tested it and kept using it I could have had random crashes and randomly corrupted files.

 

Running a test in an automated fashion every so often would help avoid that.

Handy utility! Thank you :).

 

Question: Will this be able to identify issues with ECC Ram as well?

20 hours ago, DoeBoye said:

Handy utility! Thank you :).

 

Question: Will this be able to identify issues with ECC Ram as well?

 

Not directly (unless uncorrectable), but ECC corrections would be displayed in the system log.

Tried to run this utility a few times with the same code 1 error. Any idea if this is a legitimate memory error or am I doing something wrong? 
 

I have been having issues with the OS crashing, but no errors have been found on memtest after 10 passes. 
 

Edited March 21, 20251 yr by Galax

6 minutes ago, Galax said:

Tried to run this utility a few times with the same code 1 error. Any idea if this is a legitimate memory error or am I doing something wrong? 
 

I have been having issues with the OS crashing, but no errors have been found on memtest after 10 passes. 
 

 

Do not put a space between the number and unit, write 5G.

I had a recent kernel hang and as I'm not able to physically able to get to the device for a few days, stumbled across this as a way of initial investigation. Great tool, thanks!

[16.11.2025 04:34:40 GMT] /usr/bin/memtester 28G 2
=============================================================================================

The detailed error output is enabled, watch this panel for any occurring errors.

FAILURE: possible bad address line at offset 0x000000036eef1328.
FAILURE: 0xffffffffff7fffff != 0xfffffbffff7fffff at offset 0x00000000fb7b35d8.

What is interesting is the timing of this, I installed this new NAS a month ago in my comms room downstairs and it's been running fine. On Wednesday I moved it up to my loft as my sparky came and extended my network up there - it's cooler up there (although at idle the system runs at 18-25 degrees C and the drives around the same), but also the power socket has a max limit of 5A. Hanging off the power socket are the UPS (Eaton 3S850B 3S Gen2 Desktop UPS Uninterruptible Power Supply (510W/850VA)) which connects my 2.5GbE switch and then this NAS. Is this just a red herring and just coincidental timing?

In the meantime, should I just power off the NAS till I get back and raise a warranty replacement request in the meantime? Then there's the question around the timing of the relocation, is it possible the lower temps and/or the power running to it are what's maybe caused the issue. If so, a warranty replacement may not help and the same issue might reoccur if it's either temp or power that is/has caused the issue.

12 hours ago, flashback said:

I had a recent kernel hang and as I'm not able to physically able to get to the device for a few days, stumbled across this as a way of initial investigation. Great tool, thanks!

[16.11.2025 04:34:40 GMT] /usr/bin/memtester 28G 2
=============================================================================================

The detailed error output is enabled, watch this panel for any occurring errors.

FAILURE: possible bad address line at offset 0x000000036eef1328.
FAILURE: 0xffffffffff7fffff != 0xfffffbffff7fffff at offset 0x00000000fb7b35d8.

What is interesting is the timing of this, I installed this new NAS a month ago in my comms room downstairs and it's been running fine. On Wednesday I moved it up to my loft as my sparky came and extended my network up there - it's cooler up there (although at idle the system runs at 18-25 degrees C and the drives around the same), but also the power socket has a max limit of 5A. Hanging off the power socket are the UPS (Eaton 3S850B 3S Gen2 Desktop UPS Uninterruptible Power Supply (510W/850VA)) which connects my 2.5GbE switch and then this NAS. Is this just a red herring and just coincidental timing?

In the meantime, should I just power off the NAS till I get back and raise a warranty replacement request in the meantime? Then there's the question around the timing of the relocation, is it possible the lower temps and/or the power running to it are what's maybe caused the issue. If so, a warranty replacement may not help and the same issue might reoccur if it's either temp or power that is/has caused the issue.

Thanks for the feedback, I'd say this is definitely a coincidence and I'd shutdown the system and RMA the memory, good you caught it early-ish before it caused more problems.