Re: Disk test of RAID-1 takes more than 30 hours - something wrong?

janpeter1 · ‎2021-10-13

Hi

I am in the process of upgrading my RAID1-disk pair on RN314 system. Before change one of the disks I do "Disk test" from the NAS.

It takes very long time and makes me a bit worried, although from the data I see and understand the RAID-system is found healthy so far. I have now run the disk test for about 30 hours. During the disk test I put the fan on "cool" and keep the CPU temp < 50C.

A couple of days before starting the procedure I did a partial scrub in the RAID-1 system but after 10% work and 10 hours operation I decided to stop the scrub. Both disks on the RAID are completely health, anyway, what I can see.

Some facts about the RAID1-disks. It is an odd couple. Both are WD and running on 5400 rpm. The old disk is 4 TB and 64 GB cache. The relatively new disk is 8 TB 256 GB cache. (And I am in the process of exchanging the old 4 TB disk to WD Red Plus 8 TB 256 GB cache running at 7200 rpm, since WD discontinued the 8 TB model I bought about a year ago).

How long should I let the disk test go on?

Something I can check to understand that the test disk process is running well?

(And I have done the procedure to back up everything and also the very new disk about to be used has successfully passed the disk test that took 15 hours. )

Appreciate your advice

Model RN314, latest firmware 6.10.5 hotfix 1

janpeter1 · ‎2021-10-19

I just want to repor that the final resync of "empty" last half of the 8TB was completed without any problem. All disks are ok. Took 26 hours the last part.

The somewhat long time is partly due to regular but small time machine backup from one computer that I did not turn that off during resync. That backup was however done to another disk on the same NAS but took likely some CPU resources.

Thanks for your helpful support!

View solution in original post

StephenB · ‎2021-10-13

@janpeter1 wrote:

I am in the process of upgrading my RAID1-disk pair on RN314 system. Before change one of the disks I do "Disk test" from the NAS.

It takes very long time and makes me a bit worried, although from the data I see and understand the RAID-system is found healthy so far. I have now run the disk test for about 30 hours. During the disk test I put the fan on "cool" and keep the CPU temp < 50C.

How long should I let the disk test go on?

I'd let it complete, since it does sound like there could be an issue, and if a disk is causing the problem then it should be replaced first.

@janpeter1 wrote:

How long should I let the disk test go on?

Something I can check to understand that the test disk process is running well?

You could check the log page (chosing both disk and volume as filters).

If ssh is enabled, you could also try smartctl -x /dev/sda and smartctl -x /dev/sdb

janpeter1 · ‎2021-10-13

Hi agaain,

The disk test completed after 33 hours and 3 min (on my RND314 bought 2015).

The two disk are

- WD 4 TB - 64 MB cache - 5400 rpm - WD40EFRX - bought 2015 August

- WD 8 TB - 256 MB cache - 5400 rpm - WD80EFAX - bought 2021 February

The RAID-1 had only 10% left now (of the 4TB) and high time to upgrade the 4TB disk.

After the test the overview showed the RAID-1 still as healthy

And in Perfromance view hoovering over the green knobs for each disk

I see that there an no error in any of the two disks.

Downloaded the System_log that contains a lot. I guess the file disk_info.log is of main interest.

That file shows also no errors. Here is a btrfs.log that is not too long but does not seem

to contain something worrisome. Cannot find a date here though.

I do tno think I have ssh-enabled. Many years ago on other computers I used ssh and no it is powerful tool.

Have rudimentrary Linux skills and could manage simple check ups if needed.

But what I see make me less worried now. Still the very long disk test bring questiona marks also for you, right?

One reason might be that the disk was that full, or? Still I though the disk test was without concern of what was on the disk.

Another reason for the long time might be that has been a poorly matched RAID-1 pair of disk, but not sure.

I wait for your response, before I exchange the 4 TB disk to the newly bought and tested 8TB WD80EFBX disk (7200 rpm though).

And I plant to do that echange of disk as a hot-swap, based on earlier recommendations here on the forum.

Thanks!

StephenB · ‎2021-10-14

@janpeter1 wrote:

The disk test completed after 33 hours and 3 min (on my RND314 bought 2015).

The two disk are

- WD 4 TB - 64 MB cache - 5400 rpm - WD40EFRX - bought 2015 August

- WD 8 TB - 256 MB cache - 5400 rpm - WD80EFAX - bought 2021 February

Still the very long disk test brings question marks also for you, right?

My RN524 with 4 WD80EFAX drives completes the disk test in about 21 hours. I'd expected the RN314 to complete more quickly, especially with fewer disks. But the test is run in the background, so the running time would depend on what else the NAS is doing.

@janpeter1 wrote:

I wait for your response, before I exchange the 4 TB disk to the newly bought and tested 8TB WD80EFBX disk (7200 rpm though).

And I plant to do that echange of disk as a hot-swap, based on earlier recommendations here on the forum.

... And I have done the procedure to back up everything

Since you have an up to date backup, it is safe to proceed. I agree hot-swapping is the right method.

janpeter1 · ‎2021-10-14

Hi again,

I have moved the new disk to the RAID-1 (after DESTROYING the temporary volume I made for test purpose on the new disk).

The resync goes on and now after about 2 percen done the time estimate seems to stabilize around 15 hours. Seems ok.

The fan is also turned up to "Cool".

The NAS does not do much this week, but occasionally do incremental backup of my or my wifes Macs.

So the work-load I do not think explain the long time to do Disk test.

For Netgear it would be good if they could include a progress bar in Disk test as they have for Resync.

At least you see that something happens and there is some progress, and at a later stage i guess the reamining time estimate

becomes rather good.

Within a few months I think I will buy another 7200 rpm disk so RAID-1 have the same disks.

The 5400 rpm disk I will have use for to take backups for our Macs here.

Can you say something about the noise level of 7200 rpm compared to 5400 rpm disks?

Noticeable, when you have just a RAID-1?

StephenB · ‎2021-10-14

@janpeter1 wrote:

Can you say something about the noise level of 7200 rpm compared to 5400 rpm disks?

Noticeable, when you have just a RAID-1?

WD80EFBX: 27 dBA Idle/29 dBA Seek

WD80EFAX: 27 dBA Idle/29 dBA Seek

WD40EFRX: 25 dBA Idle/28 dBA Seek.

So the WD80EFBX model has the same noise spec as the WD80EFAX you already have - and a very slight increase in the noise spec over the WD40EFRX.

Note the WD80EFAX you already have is 7200 rpm, not 5400. There is a footnote in the datasheet ( https://documents.westerndigital.com/content/dam/doc-library/en_us/assets/public/western-digital/pro... 😞 Actual Spindle Motor Rotational Speed for this model is 7200 RPM; though ID Device may report 5400 to reflect previous Performance Class designation.

Not sure what to make of that. There is some press on it here https://www.extremetech.com/extreme/317327-western-digital-changes-its-reported-drive-speeds-to-refl...

janpeter1 · ‎2021-10-14

Interesting! Means I may relax and wait a bit to replace my third 4 TB disk. But, cast some shadow over WD sort of. / Jan Peter

janpeter1 · ‎2021-10-15

Hi,

The Resyn process of the RAID-1 has come to a point where it changed from Degraded to Redundatn. It took 16 hours instead fo 15 which as an early estimate that I made from a number of such such estimates coming up with large spread. I think it is good.

What suprise me is that in now continue to Resync remaining 4 TB.

it takes another 15 hours or so I guess. And here is what I understand no material of interesst. Remember that before I had 8 TB + 4 TB and now I have 8 TB + 8 TB.

I would appreaciate some brief explanation of why this last part of Resync is needed.

Or if have done something wrong, some information not given, or so?

But I have no problem to just let it go on.

StephenB · ‎2021-10-15

@janpeter1 wrote:

Remember that before I had 8 TB + 4 TB and now I have 8 TB + 8 TB.

I would appreciate some brief explanation of why this last part of Resync is needed.

When you had 8TB+4TB, your system had a single 4 TB RAID-1 group (using both disks). The remaining 4 TB on your 8 TB disk wasn't used at all.

When you upgraded the 4 TB drive, the system creates a second 4 TB RAID-1 group to use the rest of the space. This is concatenated with the first RAID group, so you still have a single 8 TB volume.

But the two RAID groups still both need to be resynced. The first phase is to rebuild the original RAID-1 group (copying data from the original 8 TB drive to the new one). You won't get any more space in this phase.

When that is complete, the second RAID group is created, and that is also synced. That is when you will see the volume expand.

Note that if you ever do a factory default, the system would just create a single 8-TB RAID-1 group. This more complicated process is just used for vertical expansion.

janpeter1 · ‎2021-10-19

I just want to repor that the final resync of "empty" last half of the 8TB was completed without any problem. All disks are ok. Took 26 hours the last part.

The somewhat long time is partly due to regular but small time machine backup from one computer that I did not turn that off during resync. That backup was however done to another disk on the same NAS but took likely some CPU resources.

Thanks for your helpful support!