DiskSpeed, hdd/ssd benchmarking (unRAID 6+), version 2.10.8


Recommended Posts

3 hours ago, Zonediver said:

 

What shall i add and were? I am not a Docker-Expert...

 

Click on the icon as if you were going to launch the app but instead click "Edit"

image.png.68c35542847097a048e161f30e95d93f.png

 

At the bottom of the page, click on

image.png.75f738857da71c0d5c0c6580991103be.png

 

Set to the following:

image.png.a0acde630d00e0b7ef61751b37f999ee.png

 

Note on Access Mode: Since the docker app runs in Privileged mode, it doesn't matter if it's Read Only or Read/Write, but might as well select Read/Write as future enhancements to allow the app to run outside of Privileged mode will require this.

 

Save the new setting, then click on the "Apply" button for the Docker (not the Save button).

 

Then launch DiskSpeed as per normal.

 

  • Thanks 1
Link to comment
19 hours ago, jbartlett said:

Save the new setting, then click on the "Apply" button for the Docker (not the Save button).

Then launch DiskSpeed as per normal.

 

This did the trick - thanks for your detailed help.

The question is now: Why was this Docker installed at the other path?

And why did it work before?

I installed it over the Apps-function - so i did not set this path - the install-process did this...

 

EDIT: There is something strange at the write-speeds...

The Samsung SSD 970 PRO is far more faster then 397MB/s.

When i transfere data over my 10GBit NIC, i reach up to 1189MB/s constantly...

The constant write speed of this SSD is 2300MB/s.

 

grafik.thumb.png.09bcc880c3fdafdcf774701815501907.png

 

Edited by Zonediver
  • Like 1
Link to comment
10 hours ago, Zonediver said:

 

This did the trick - thanks for your detailed help.

The question is now: Why was this Docker installed at the other path?

And why did it work before?

I installed it over the Apps-function - so i did not set this path - the install-process did this...

 

EDIT: There is something strange at the write-speeds...

The Samsung SSD 970 PRO is far more faster then 397MB/s.

When i transfere data over my 10GBit NIC, i reach up to 1189MB/s constantly...

 

One thing to keep in mind is that when you're copying files, it takes advantage of both system and device cache and DiskSpeed specifically requests that "dd" to not utilize any system caches - though that doesn't mean the device has to honor the request.

 

The 970 Pro is supposed to get 2.700 MB/S on write so something's strange going on. View the drive and click the Benchmark button. While it's benchmarking, look at the "Click on a drive label to hide or show it." line under the graph and click just to the right of the period at the end - this will toggle the visibility of the hidden iframes doing the actual work. When it's writing the files, take note of lines like this and reply with what you got.

dd if=/dev/urandom of=/mnt/UNRAID/cache/DiskSpeedTestFile01.junk bs=131072 count=16384 oflag=direct conv=noerror status=progress
231 MB/Sec (9260 ms)

 

@jsmontague - please use the same step to show the hidden iframe in your case, it will display an error message. Please copy that into a reply (don't need the long java stack trace at the end).

Link to comment

Just following on though, this is something I don't think I have seen before when running diskspeed:

 

image.png.db9611b4dc95c2029301974a9e703bb5.png

 

The drive is an old 2TB platter type, is it correct that it is trimming? Also it looks to be stuck on 0%. This drive has not benchmarked, either. 

Edit: this drive seems to be getting detected as a SSD...

image.png.dd779f2c8f973cfba24410b24098f0c6.png

Edited by anthropoidape
Link to comment
On 12/27/2022 at 7:07 AM, jbartlett said:

@d3m3zs@jsmontague@Zonediver@emb531@youngfish@deepinder27@mayberts

 

Please update DiskSpeed to 2.10.2. I found an issue with detecting partitions on nvme drives and added additional error trapping with logging during the scan process. If you continue to have any issues, please create a new Debug File from inside the app or by using:

http://[IP]:18888/isolated/CreateDebugInfo.cfm

 

The new debug file will also contain a html file of the hardware scan so I'll be able to see exactly what you saw if there was an issue during the scan process.

 

Keep in mind that SSD's and NVME drives can't be benchmarked if they are not operating as solitary drives - that means no RAID, multi-pool BTRFS setups, etc. If you want to benchmark such drives, it has to be done prior to adding them to a multi-drive setup.

So I can benchmark my xfs nvme array now, thank you. I'm currious as to why I can no longer benchmark my btrfs nvme's?  I can on diskspeed version :2.9.7

Link to comment
13 hours ago, jbartlett said:

The 970 Pro is supposed to get 2.700 MB/S on write so something's strange going on. View the drive and click the Benchmark button. While it's benchmarking, look at the "Click on a drive label to hide or show it." line under the graph and click just to the right of the period at the end - this will toggle the visibility of the hidden iframes doing the actual work. When it's writing the files, take note of lines like this and reply with what you got.

 

...there is nothing to "click"... just hide or show (r/w) - but no further infos there...

 

grafik.png

 

grafik.thumb.png.abacfc080e4ca837f78b28fe315aac3a.png

Edited by Zonediver
Link to comment
11 hours ago, Zonediver said:

 

...there is nothing to "click"... just hide or show (r/w) - but no further infos there...

 

It's the "nothing" you click on. :) - in this orange area just to the right of the period. Your mouse pointer should still be in the text highlight pointer and not a mouse pointer (on Windows)

 

image.png.a1b6d826cd359196f5dba6d31e833b06.png <= Click in orange area

Link to comment
13 hours ago, jbartlett said:

 

It's the "nothing" you click on. :) - in this orange area just to the right of the period. Your mouse pointer should still be in the text highlight pointer and not a mouse pointer (on Windows)

 

image.png.a1b6d826cd359196f5dba6d31e833b06.png <= Click in orange area

 

Well... maybe i am too sily, but there is no "orange area"... nowere... sorry

 

EDIT: Now i got it 🤣

These are the values of my SSD 970 Pro from Samsung:

grafik.png.1cf2b0caf833a28742ff867fe9ec13ee.png

Edited by Zonediver
Link to comment
10:23:16 Spinning up sdu (6TB)
10:23:16 Performing random seek tests
10:23:26 Performing sequential seek tests
10:23:38 Performing drive latency tests
Random Seek: 101
SequentialSeek: 155
DriveLatency: 8229412
Lucee 5.3.10.97 Error (application)
Message	Error invoking external process
Detail	fstrim: /mnt/UNRAID/disk14: the discard operation is not supported
Stacktrace	The Error Occurred in
/var/www/BenchmarkDriveController.cfm: line 131
129:
130: <CFFILE action="write" file="#OutDir#/trim.sh" mode="766" output="fstrim #MountPoint#" addnewline="NO">
131: <CFEXECUTE name="#OutDir#/trim.sh" timeout="3060" />
132: <CFSET WriteSpeedList="">
133: <CFLOOP index="i" from="1" to="#SSDBenchmarkTestFiles#">

 

Was able to get one to benchmark and this is the output as it gets stuck at trimming.

Link to comment

Does anyone know why although DiskSpeed can see all my drives in the left hand column, it only thinks there are five disks in the benckmark. If I tey and benchmark one of the other drives (e.g Port 6:Disk 6 (sdf)) DiskSpeed seems to hang before completely drawing the graph at the start, but works fine on any of the drives it "knows about".GrahamWatson_20230106_115511.tar.gz

 

FYI, I also didn't have the UNRAID attribute in my docker install.  I added it but it doesn't seem to have made any difference.

 

Screen Shot 2023-01-06 at 11.50.02 AM.png

 

Screen Shot 2023-01-06 at 11.52.45 AM.png

 

One more thing - when I benchmark the controller, it does include the drives that it won't benchmark when I benchmark the disks...

 

Screen Shot 2023-01-06 at 1.46.46 PM.png

Edited by GrahamTWatson
Link to comment
On 1/5/2023 at 10:25 AM, jsmontague said:
10:23:16 Spinning up sdu (6TB)
10:23:16 Performing random seek tests
10:23:26 Performing sequential seek tests
10:23:38 Performing drive latency tests
Random Seek: 101
SequentialSeek: 155
DriveLatency: 8229412
Lucee 5.3.10.97 Error (application)
Message	Error invoking external process
Detail	fstrim: /mnt/UNRAID/disk14: the discard operation is not supported
Stacktrace	The Error Occurred in
/var/www/BenchmarkDriveController.cfm: line 131
129:
130: <CFFILE action="write" file="#OutDir#/trim.sh" mode="766" output="fstrim #MountPoint#" addnewline="NO">
131: <CFEXECUTE name="#OutDir#/trim.sh" timeout="3060" />
132: <CFSET WriteSpeedList="">
133: <CFLOOP index="i" from="1" to="#SSDBenchmarkTestFiles#">

 

Was able to get one to benchmark and this is the output as it gets stuck at trimming.

@jbartlett anything else you need from me? 

 

Link to comment

Hello.

Thanks for the cool program. I recently purchased a new drive and thanks to Diskspeed I found that it was faulty. I did a preclear and noticed the times were a little off in the last part of the drive. I used diskspeed to run a test and I found that around the 7TB mark to speeds dropped dramatically - not smart errors were detected or anything but the benchmark speed dropped to about 10 MB/s. One of the things that I did with the test was I changed a setting to benchmark the drive every 1% instead of the default 10%. Now I can't find that setting. Can you please show me where it is?

 

Thanks

Link to comment

Sorry for the late replies, been a busy couple of weeks for me.

 

I found out the problem with the Maiwo KT015 PCIe card with NVMe & SATA ports is a bug on the card itself. Issuing multiple "nvme show-regs" commands in a row will cause the system to hang without any mentions of issues in the syslog.

 

Signs of the buggy card is the all F's in the pmrmsc field

 

/usr/sbin/nvme show-regs /dev/nvme1n1
cap     : 100030f00103ff
version : 10300
cc      : 460001
csts    : 1
nssr    : 0
intms   : 0
intmc   : 0
aqa     : 1f001f
asq     : 105977000
acq     : 105978000
cmbloc  : 0
cmbsz   : 0
bpinfo  : 0
bprsel  : 0
bpmbl   : 0
cmbmsc  : 0
cmbsts  : 0
pmrcap  : 0
pmrctl  : 0
pmrsts  : 0
pmrebs  : 0
pmrswtp : 0
pmrmsc  : ffffffffffffffff

 

Repeat it a few times and every value returns all f's. Then if done again, system locks up.

 

Looking into another way to get the data from that list.

 

In other news, I suspect I know what's causing the slow write benchmarks on SSD/nvme drives but haven't tested that theory yet. I haven't yet looked into why it's trying to trim a spinner. That's just silly. I hope to get to it this weekend.

 

@ToastinJam - You can select a 1% test range from the "Benchmark Drives" button on the main page. Clicking the banner will take you there.

 

@GrahamTWatson - The benchmark graph on the main page only displays drives you've done benchmarks on. If you've purged the settings, it erases past benchmarks too though you can recover them if you uploaded your benchmark data. To do so, click on any drive and click the Maintain Benchmark button - though you have to have at least one benchmark already stored for it to display.

  • Thanks 1
Link to comment

I found it kept saying it couldnt test any disk that is not part of the array but then after 5mins it seems to have found the nvme drives and is testing them. Will see if it finds the other sata ssds.

 

Don't think these new speeds are quite right? We can see the previous method below. A cool docker non the less.

 

839940672_Screenshot2023-01-14at08_38_00.thumb.png.7eac65e521168a38f92652f871124c56.png

Edited by dopeytree
Link to comment
On 1/1/2023 at 3:30 AM, mayberts said:

So I can benchmark my xfs nvme array now, thank you. I'm currious as to why I can no longer benchmark my btrfs nvme's?  I can on diskspeed version :2.9.7

 

Same.

 

BTRFS NVME solo drive with a single VM on it says:

 

"Unable to benchmark for the following reason
* No mounted partitions were found.
You will need to restart the DiskSpeed docker after making changes to mounted drives for changes to take effect."

 

Been working with it all day so I know it has a fs. :)

 

Link to comment

Guys for anyone having the issue of some HDD models recognized as SSD, I found a temporary fix: rollback to version 2.10.

How to do that? Just edit the docker settings and set this field as follows:

image.thumb.png.b99fb299cb59126c8f91b340ed5e3b15.png

 

if the issue still persists, just remove the docker, delete the following folder "/mnt/user/appdata/DiskSpeed/" or wherever you've saved the app data of this docker, then when you reinstall the docker, set the repository field as "jbartlett777/diskspeed:2.10".

 

In the future, when the bug will be fixed you'll just have to edit the docker again and set the repository to pull the latest version as follows: "jbartlett777/diskspeed".

Hope this will help! ;) 

  • Thanks 1
Link to comment
  • jbartlett changed the title to DiskSpeed, hdd/ssd benchmarking (unRAID 6+), version 2.10.8

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.