5 hours ago, iceperson said:

When trying to benchmark my drives I get this warning:

WARNING: You have less than 4 CPUs available (0 total). The Solid State benchmark requires 4 CPU threads in order to get an accurate benchmark. Solid State Benchmarking is not available.

 

I'm on an 8700k, 6 cores 12 threads. I have Enable VMs set to 'No' in unraid since I don't run any VMs, although I did configured the BIOS virtualization for VMs and test to make sure everything is configured to work if I want to run any in the future. No other containers are running, no CPU pinning anywhere. I did try to pin 4, 5, and 6 CPUs to DiskSpeed but still get the warning and unable to benchmark SSDs.

 

I'm seeing the same on my dev box. Did you upgrade to Unraid 6.12 RC? Looks like the OS path to get the utilized CPU's from inside the Docker app changed.

Edited March 31, 20233 yr by jbartlett

1 hour ago, jbartlett said:

 

I'm seeing the same on my dev box. Did you upgrade to Unraid 6.12 RC? Looks like the OS path to get the utilized CPU's from inside the Docker app changed.

 

Oh snap. I did. Sorry.

4 hours ago, iceperson said:

 

Oh snap. I did. Sorry.

 

No, I'm glad you did! You caught this and brought it to my attention. Now I can fix it before things go out of RC. Thanks!

Hi,
I got two unbenchmarkable NVMEs:

NVMEs: 2x transcend PCIe SSD 220S (https://de.transcend-info.com/products/No-991) in a btrfs raid1-pool
Board: ASRock H470M-ITX/ac (https://www.asrock.com/MB/Intel/H470M-ITXac/)
Unraid: Version 6.11.5
Diagnostic send per Mail.

7 hours ago, sonic6 said:

Hi,
I got two unbenchmarkable NVMEs:

NVMEs: 2x transcend PCIe SSD 220S (https://de.transcend-info.com/products/No-991) in a btrfs raid1-pool
Board: ASRock H470M-ITX/ac (https://www.asrock.com/MB/Intel/H470M-ITXac/)
Unraid: Version 6.11.5
Diagnostic send per Mail.

 

DiskSpeed version 2 will not benchmark multiple-drive pools, it's purpose is benchmarking single drives. I'm thinking of adding support for it in version 3 which is currently in development.

7 hours ago, jbartlett said:

DiskSpeed version 2 will not benchmark multiple-drive pools, it's purpose is benchmarking single drives.

Interessting. I upgraded my btrfs-raid1 cache pool yesterday from two SATA mx500 SSDs to this PCIe 220S NVMe SSDs. The mx500 SATA drives (part of pool) where benchmarkable.

@jbartlett

Hi, I just downloaded diskspeed, really nice to be able to benchmark the drives inside unraid, thank you for this container.

 

I was wondering if you could take a look at this.

While benchmarking my Crucial nvme, I get this as a end result.

As you can see, the " Avg Read / Write Speed " text, is clumbed together, and very hard to see.

The MB/Sec on the left doesn't seem correct, isn't it suppose to say GB/Sec for the nvme ?

 

I would also like to request a way to be able to get back the older results again to get to this page.

 

 

8 hours ago, Zidichy said:

I was wondering if you could take a look at this.

While benchmarking my Crucial nvme, I get this as a end result.

 

It looks like the scale on the left is incorrect, that it's actually GB instead of MB - it indicates GB in the top summary and on the right side of the graph. If you could send me a Debug file (click Debug link inside app, bottom left corner) to [email protected], I can take a look. The big spike in the middle (couldn't help but snicker) looks to be some kind of cache related issue, not sure what would cause that.

 

You can download an older version of the app by specifying the version number in the Docker resource tag.

For example: jbartlett777/diskspeed:2.9.7

 

I can tell you with absolute certainly that the SSD benchmarks on the older versions are wrong. It only has one read/write stream which is not enough to saturate the IO buffer in all cases. The latest version creates 4 threads by default, each thread working on a different file.

Thank you, I will send the debug file.

On 3/31/2023 at 8:57 PM, jbartlett said:

 

I'm seeing the same on my dev box. Did you upgrade to Unraid 6.12 RC? Looks like the OS path to get the utilized CPU's from inside the Docker app changed.

Hello

i have also the 6.12rc3 and the same problem
WARNING: You have less than 4 CPUs available (0 total). (AMD Ryzen 3600)

is there a way to bypass this?
Or will that be fixed in the next version?

thanks a lot

I feel like I must be missing something obvious, but I can't seem to select my SSD cache drives for testing. They are listed, but grayed out/unselectable. I did add the path value specified in the FAQ.

 

14 minutes ago, crafty35a said:

I feel like I must be missing something obvious, but I can't seem to select my SSD cache drives for testing. They are listed, but grayed out/unselectable. I did add the path value specified in the FAQ.

 

i think you have also this 'problem'

see down below your side

16 minutes ago, luckyheiko said:

i think you have also this 'problem'

 

Thank you, but I don't see that warning message anywhere (and also I have 6 CPU cores)

On 4/16/2023 at 12:10 PM, luckyheiko said:

is there a way to bypass this?
Or will that be fixed in the next version?

thanks a lot

 

There is no easy way to bypass this. The files that give the way the CPU's are represented inside the Docker image has changed paths. The next patch will update the file references.

 

8 hours ago, crafty35a said:

I feel like I must be missing something obvious, but I can't seem to select my SSD cache drives for testing. They are listed, but grayed out/unselectable. I did add the path value specified in the FAQ.

 

You have two drives in your cache pool. DiskSpeed 2 will not benchmark multi-drive arrays/pools/raid as it's designed to test single drives. That'll be a feature in DiskSpeed 3.

14 hours ago, jbartlett said:

You have two drives in your cache pool. DiskSpeed 2 will not benchmark multi-drive arrays/pools/raid as it's designed to test single drives. That'll be a feature in DiskSpeed 3.

 

But, why i was able to benchmark my two SATA SSDs from my btrfs-raid1 pool in the past?

4 hours ago, sonic6 said:

But, why i was able to benchmark my two SATA SSDs from my btrfs-raid1 pool in the past?

 

Most likely the methods I programmed to detect multi-drive pools didn’t catch your specific scenario or you were on an older version of the app. One of my systems has a btrfs-raid1 cache pool and it detected it and stopped it from benchmarking.

I've got version 2.10.5 ready to upload which fixes the SSD benchmarking under UNRAID 6.12 but for some reason the app doesn't want to push partial content to the browser and instead waits until it's done before doing. For example, if you benchmark a drive, you will see a white screen until it's done. I've had this happen before but I don't recall what the cause was or how to resolve it. Time to get Googleing!

 

Not only did the new version of Docker changed how it represents the internal working state of the virtualized hardware from inside the Docker container, it's not always the same from one run to the next. But I've identified what I believe to be all the locations needed to look at for the info so we should be good to go once I solve the other issue above.

Edited April 24, 20233 yr by jbartlett

Version 2.10.5 deployed.

• Updated to support Docker 20.10.x for SSD benchmarking. The files that indicate the CPU configuration inside the Docker containers changed.
• Clarified the missing mountpoint to indicate it's missing in Docker such as "Docker volume mount not detected"
• Added caching detection logic that will display a warning on the hardware scan if the HTML is being intercepted by a proxy or 3rd party app such as Acronis Cyber Protect Home Office.

Hi all - Having an issue with benchmarking. Every time I "Start Drive Benchmark" or "Benchmark Drive" I see the below blank screen. No other issues or warnings noted. Any ideas on what I can try for next steps?

 

7 hours ago, sidens said:

Hi all - Having an issue with benchmarking. Every time I "Start Drive Benchmark" or "Benchmark Drive" I see the below blank screen. No other issues or warnings noted. Any ideas on what I can try for next steps?

 

Click on the period at the end of "hide or show it" to display the hidden iframes that are doing the actual work to see the progress or an error message.

May I ask carefully when version 3 is gonna be released?

Looking forward to test my usb drives - is there maybe a beta channel?

 

And thank you very much for the development, I really appreciate this!

Best regards

Hello!

I have some issue testing my Crucial P5 SSD 2 TB.

I have 9 drives in total in my configuration, two of them being SSD, an old SATA SSD and the Crucial P5.

Everything works well until I test the P5... it starts well by Performing random seek tests, then it does the latency tests  but after that it stucks on Trimming drive... the progress bar doesn't show anything (neither the graph actually) and I am unable to tell if it is actually doing something or not. I assume something goes wrong as the temperature of the SSD goes back to normal....

 

thanks in advance for your help!

Marco

 

 

On 5/2/2023 at 11:10 AM, Marcco said:

Hello!

I have some issue testing my Crucial P5 SSD 2 TB.

I have 9 drives in total in my configuration, two of them being SSD, an old SATA SSD and the Crucial P5.

Everything works well until I test the P5... it starts well by Performing random seek tests, then it does the latency tests  but after that it stucks on Trimming drive... the progress bar doesn't show anything (neither the graph actually) and I am unable to tell if it is actually doing something or not. I assume something goes wrong as the temperature of the SSD goes back to normal....

 

thanks in advance for your help!

Marco

 

Click on the period (just to the right of it) at the end of "hide or show it" line when benchmarking to display the hidden iframes that are doing the actual work. If there was an error during the trim process, it will show up there.

On 4/12/2023 at 10:20 AM, Zidichy said:

@jbartlett

Hi, I just downloaded diskspeed, really nice to be able to benchmark the drives inside unraid, thank you for this container.

 

I was wondering if you could take a look at this.

While benchmarking my Crucial nvme, I get this as a end result.

As you can see, the " Avg Read / Write Speed " text, is clumbed together, and very hard to see.

The MB/Sec on the left doesn't seem correct, isn't it suppose to say GB/Sec for the nvme ?

 

I am also getting similar results on a brand new top-of-the-line Seagate FireCuda 530 1TB nvme drive.

 

Is it possible not enough threads are being used in order to saturate the drive read/write capacity?

 

I'm trying to diagnose why I'm suddenly no longer getting a sustained 10Gb/s of copy throughput from my machine to my Unraid machine (both have 10Gb NICs and are connected via a 10Gb switch).  At one point I was able to copy movies at full network capacity, and iPerf 3 confirms that the network is indeed functioning at full speed. That leaves the disk (or OS) as the primary suspect.

 

So I just replaced the previous 500GB FireCuda 530 with this new 1TB (the read/write speed specs on the 1TB are slightly better) and nothing changed in terms of real world copy speeds. However, the benchmarks indicate a minimum of 3.8 GB/sec of write speed. (BTW, the average write speed line looks incorrect in my results).

 

When I'm copying movies, they start out at around 2.3GB/s and eventually grind down to around 600MB/s. It's bizarre and I have no idea what's causing it.

2 hours ago, JG76 said:

Is it possible not enough threads are being used in order to saturate the drive read/write capacity?

 

Interesting thought. I'll look into adding an option to specify the number of threads.