DiskSpeed, hard drive benchmarking (unRAID 6+), version 2.9.2


601 posts in this topic Last Reply

Recommended Posts

On 8/14/2020 at 2:00 PM, johnnie.black said:

Those are SMR, they will only show the correct performance after being written completely at least once.

 

On 8/14/2020 at 9:52 PM, jbartlett said:

If you notice how the drives have a flat line? That's what the program looks for to determining if bandwidth was capped - or if it seems the drive can output data faster than the drive controller or connection can handle.

 

As per tee-tee jorge's comment, run a pre-clear on those drives and you should get graphs that match the trend on the Hard Drive Database

 

Thanks for the info. I wasn't aware of needing to pre-clear first. Anyway I've done that now and the graph looks the same as the Parity drive, so all good.

Link to post
  • 1 month later...
  • Replies 600
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Popular Posts

This Docker Application will let you view your storage controllers & the drives attached to them and perform Benchmarks on both. Controller Benchmarks helps to identify if the drives attached to i

I'm taking this application out of BETA status. Version 2.0 has been released.   Release 2.0 Added progress bars to the drive benchmarking Rewrote the Controller Benchmark to be

I just noticed a trend in the file fragmentation. It seems the OS has the tendency to break the file up into chunks of a set size. I found this really strange because, well, why do it at all? Noticing

Posted Images

Hi @jbartlett,

 

Thanks for putting the docker together, I've got it installed on 2x unRAID boxes and 1 works fine but the other does not.

 

Motherboard: Asus B85M-E

HBA: H310 (flashed to LSI HBA mode)

 

The SSD's where originally connected to the HBA but I've moved them to the onboard SATA controller and I get the same error message. I think the SSD make/model being identical up until the last 3 digits might be the issue?

 

Any help would be appreciated :)

image.png

image.png

image.png

Link to post

I see the issue, it's a typo. It's passing a numeric position to the ListFirst function at the end of the line which doesn't take a numeric position. It's too late for me to push a fix tonight but I'll take care of it tomorrow.

Link to post
  • 3 weeks later...

As a (quite genuinely humble in the light of the expertise on display here) newbie to this topic and to UnRAID in general, may I make a plea on behalf of my old and now deceased chemistry master, Mr Creighton, who would regularly throw chalk with high speed and deadly accuracy at any pupil who omitted or misapplied units.

 

He is turning in his grave at the use of "M" on the Y-axis and "G" on the X-axis for these drive speed charts.

 

-- 

Chris

 

 

Edited by bidmead
reply notification not ticked
Link to post
20 minutes ago, bidmead said:

He is turning in his grave at the use of "M" on the Y-axis and "G" on the X-axis for these drive speed charts.

Why? What is your interpretation of the chart? What do you think each axis represents?

Link to post

My interpretation probably isn't very relevant. It's Mr Creighton and his flying chalk we should be worrying about...

 

More seriously, a chart like this is supposed to convey information with a degree of precision. It's not a Rorschach Inkblot Test.

 

-- 

Chris

Link to post
3 minutes ago, bidmead said:

My interpretation probably isn't very relevant.

It's relevant to understand why you think it should be different. I suspect Mr Creighton would have no problem with it if he understood what

2 hours ago, trurl said:

each axis represents

What do you think M on the Y-axis means and what do you thing G (or T) on the X-axis means?

5 minutes ago, bidmead said:

a chart like this is supposed to convey information with a degree of precision.

It does convey information with an appropriate degree of precision.

 

Link to post

I didn't come here to pick a fight. But happy to be piled on by the occupants here if that's the style of discourse. It's a kind of welcome, I suppose.

 

As an IT journalist I'd never get away with using T on its own to denote anything other than a metric tonne (or perhaps a tablespoon). There are well-developed conventions for describing MB/s (or Mb/s---which is it?) and TB (or Tb or TiB---which is it?) and the Internet seems to confirm that there's enough regular adherence to these conventions in drive speed charts to justify the stinging flightpath of Mr Creighton's chalk in the direction of fuzzy units.

 

John Bartlett's done some really excellent work here. Are we saying it's perfect and beyond refinement?

 

-- 

Chris

 

 

Link to post
1 hour ago, bidmead said:

I didn't come here to pick a fight. But happy to be piled on by the occupants here if that's the style of discourse. It's a kind of welcome, I suppose.

 

I dont see anyone piling on. As an outside observer, I see a single question was asked. I do not see any answers provided to the question. In order to have improvements, one has to know what and how it can be improved.

 

May I suggest as a neutral observer, that you stop reading intent and other malicious actions into the simple single question, when they simply do not exist.

Link to post

Highcharts.com renders the charts and I'm passing in a large number and that's how it's translating the number. I can override the suffix but then it outputs the entire number. Highcharts understands numbers, it doesn't understand drive capacity numbers.

 

That said, if you hover over a point where I do specify the value to display, it says MB and GB.

Link to post

Many thanks for that, John. I didn't know about Highcharts. That's very useful enlightenment.

 

My comments here have been largely a knee-jerk reaction conditioned by flying chalk incidents in boyhood. I apologise if I've raised a hackle or two.

 

I still think Mr Creighton has an indelibly valid point. But at the next seance I shall tell him to calm down.

 

-- 

Chris

 

Link to post

I'm getting this error in the DriveSpeed logs:

 

"ERROR","127.0.0.1-startStop-1","11/03/2020","10:33:53","","load-extension;Extension [15mcbgnnccbzf.lex] was not found at [/opt/lucee/server/lucee-server/context/extensions/installed/15mcbgnnccbzf.lex];lucee.runtime.exp.ApplicationException: Extension [15mcbgnnccbzf.lex] was not found at [/opt/lucee/server/lucee-server/context/extensions/installed/15mcbgnnccbzf.lex]

 

Can anyone throw light on this?

 

-- 

Chris

Link to post
13 hours ago, bidmead said:

I'm getting this error in the DriveSpeed logs:

 


"ERROR","127.0.0.1-startStop-1","11/03/2020","10:33:53","","load-extension;Extension [15mcbgnnccbzf.lex] was not found at [/opt/lucee/server/lucee-server/context/extensions/installed/15mcbgnnccbzf.lex];lucee.runtime.exp.ApplicationException: Extension [15mcbgnnccbzf.lex] was not found at [/opt/lucee/server/lucee-server/context/extensions/installed/15mcbgnnccbzf.lex]

 

Can anyone throw light on this?

 

-- 

Chris

I'll look into it but you can ignore it. It's a lucee extension that I'm not using but it's trying to load.

Link to post
  • 2 weeks later...

I suspect this is an issue with my system, and not with this container, but I'm hoping someone might be able to point me in the right direction here.

I've had issues with very slow writes to my array (sub 1MB/s) at first I thought it was just the type of writes, lots of very small files with another CA docker container (urbackup).  But after nuking all its shares and existing backup data and starting over thinking there was a config issue with that backup app, I decided to run this speed benchmark to check all the drives.

I have 24x 3.5" drives in the 2TB and 6TB range in the array.  When I start the bench mark it seem to keep getting "stuck" testing Disk 3 (sdp).

 

Most tests took in the range of a few min (maybe 5?) but this drives has been stuck at 36% for over half an hour.  I don't see any reads on the main tab on that drive being reported.  if I reopen the webui it goes back to the main page where I can start another benchmark.  Doing so seems to hang up in the same spot.

I've attached the diag from the main page below.  

diskspeed_20201118_182405.tar.gz

I'm afraid I have a bad drive or something and worse if thats the case that Unraid hasn't given me any warning about it.

*edit* To rule out as many background processes and disk i/o as possible I stopped all of my other Docker containers and VMs (well 2 of the 4 vms I actually paused but that shouldn't make a difference in this case I don't think).

Edited by snowmirage
Link to post
2 hours ago, snowmirage said:

Most tests took in the range of a few min (maybe 5?) but this drives has been stuck at 36% for over half an hour. 

Click on the Benchmark button on the main page and uncheck "Check all drives", set the checkbox for Disk 3, and set the checkbox for "Disable Speed Gap detection".

 

If you frequently get a "retrying" on a spot, its because the min/max speeds at that spot were over a certain threshold and it is retrying that spot again after increasing that threshold by 5mb. Setting the Disable Speed Gap button disables this retry feature. This gap between the min/max speeds is typically caused by some other process accessing the drive at the same time but it can also be caused by a bad spot. I plan on adding a cap to the number of retries.

 

The newest version of this app also displays the SMART information for the drive which by default displays only catastrophic drive health values. Take a screen shot of that and post it. If it's not displaying it, then there's something wrong with the smart report as it gets it from UNRAID and UNRAID isn't saving it. View the drive info on the unraid Main tab to get the smart values.

Link to post
11 hours ago, jbartlett said:

Click on the Benchmark button on the main page and uncheck "Check all drives", set the checkbox for Disk 3, and set the checkbox for "Disable Speed Gap detection".

 

If you frequently get a "retrying" on a spot, its because the min/max speeds at that spot were over a certain threshold and it is retrying that spot again after increasing that threshold by 5mb. Setting the Disable Speed Gap button disables this retry feature. This gap between the min/max speeds is typically caused by some other process accessing the drive at the same time but it can also be caused by a bad spot. I plan on adding a cap to the number of retries.

 

The newest version of this app also displays the SMART information for the drive which by default displays only catastrophic drive health values. Take a screen shot of that and post it. If it's not displaying it, then there's something wrong with the smart report as it gets it from UNRAID and UNRAID isn't saving it. View the drive info on the unraid Main tab to get the smart values.

That seems to have done it.  Test for that drive finished following your suggestions.

image.thumb.png.2625ad4666f8edcbd3fbb4c70c6707c9.pngimage.png.cb0f755100f0c2182df0710c770910c9.pngimage.thumb.png.7a87a3e7fcda322320a046bb70ede6ec.png

To be honest I haven't looked at a SMART report closely in years and need to go refresh my memory on what "bad" really looks like.  

Link to post

Nothing looks wrong with the drive from these reports. A future update will add heat map support where the entire drive is read and the read speeds are represented in a color range. This might help see hidden trouble spots. That update is probably a couple months away though. Right now, I'm working on adding drive allocation & fragmentation maps which the heat maps will use to locate which files are contained in found bad spots.

 

One thing you can try is to run a benchmark as given above but test the drive every 1%.

Edited by jbartlett
Link to post
3 hours ago, jbartlett said:

Nothing looks wrong with the drive from these reports. A future update will add heat map support where the entire drive is read and the read speeds are represented in a color range. This might help see hidden trouble spots. That update is probably a couple months away though. Right now, I'm working on adding drive allocation & fragmentation maps which the heat maps will use to locate which files are contained in found bad spots.

 

One thing you can try is to run a benchmark as given above but test the drive every 1%.

That would be pretty slick I look forward to testing it out.

Since your advice yesterday I've been going back through and trying to test all the drives.

I've found several that are getting that "speed gap" notice.

I've tried a few times now to test them with "Disable speed gap detection" checked and I still see reads start 90-100+ drop to 60-70 then repeated "retries" even when I have that disable speed gap detection checked.

Link to post

I did some checking on the model number. It's a White Label drive which means it's likely a specific drive built for a specific purpose and rebranded. It looks a Western Digital drive. Personally, I kinda cringed when I saw it was a WL.

 

Here's a long tread about the drives for your perusal. I only read the 1st page.

https://forums.servethehome.com/index.php?threads/review-of-white-label-4tb-drives.3559/

Link to post
  • jbartlett changed the title to DiskSpeed, hard drive benchmarking (unRAID 6+), version 2.9.2

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.