Greetings!

First of all: I'm sorry to say that, but I'm a potato in questions of IT especially Linux etc. This Unraid Server is my very first project building something like a NAS etc on my own. Without the videos by spaceinvader one, I would not have been able to get anything running here...
So please, if you have any ideas: Explain it to me like I'm five 😅


Problem:

• Disk 1 is always online and spinning. Same for the parity drive.
• Disk 2 spins down after 15mins, as ordered to
• if order Disk 1 to go to sleep, comes back up after a few seconds
• constant reading of some kilobytes from Disk 1
• Plugin FileActivity shows nothing for Disk 1 at all

 

Steps so far:

• Moved Appdata, System, Isos etc to cache; only Nextcloud-Userfiles are left on Disk 1, same for Disk 2 where it works fine- nothing changed
• stopped all dockers -  nothing changed (???)
• changed drive of Disk 1 from the Internal SSD to an external HDD - nothing changed
• installed "Dynamix Cache Directories" Plugin - nothing changed
• Made a post on that on reddit: - no solution

 

My System is a Dell Wsys 5070. Build up:

 

Array:

• Parity: 1TB HDD in an external docker-station, connected via USB
• Disk 1: 1 TB HDD external, connected by USB
• Disk 2: 500 GB external, connected by USB

Pool:

• Cache 1: 128GB SSD, internal type, connected via external USB connector

Cache 2: 128 GB M2.SSD, internal of the Dell Wyse; used to be Disk 1 before with the same problem
 

 

Video of the bahaviour, screenshot of the running Dockers and Diagnostics attached.

 

Any suggestions where to look next for the source of the activity?
If you need anything else on information, please tell me where to find that and I will be happy to provide it.

Thank you in advance for your help 😃

unsc-gatekeeper-diagnostics-20240105-1427.zip

USB not recommended for array or pools for many reasons.

I don't get this:

Disk 2 is an external HDD via USB as well and spin down works fine here.

 

Disk 1 was an internal m2.SSD before, so no USB, and it had the same issue.

Are these normal internal drives in separate external USB enclosures?

Maybe my english was not precise enough to describe the setup.
Let me try to describe it again:

Parity: a 1TB 3.5" HDD for Internal use; put in a docking Station with own power support and SATA-connectors. Connected via USB 3.0. These things where you push in a hardrive like a game in the old nintendos.

 

Disk 1: Was the internal M2.SSD, directly on the mainboard of the Dell Wyse with the same problem. Is now changed to a 1TB external HDD in a fixed enclosure; built to be an external harddrive with own power supply. Connected via USB (USB-mini to USB-A).

Disk 2: Same as disk one, but an older model, both by Seagate. 500GB HDD in fixed enclosure. Same power supply, same USB connection method.

Par = Parity

C1/2 = Cache-Drives; C2 beeing the internal M2.SSD

1/2= Array-Drives

Edited January 5, 20242 yr by Carmino

Different USB implementations can have different shortcomings, some of which can cause problems with Unraid, some more bothersome than others, including disconnects that make the array out-of-sync, and difficulty migrating disks to other setups.

 

Best results if you insist on using USB is enclosures that support UASP.

First:

With this setup, I don't have a choice other than rely on external drives.
Right now, its a mashup of different harddrives I had around without using them anymore. So my Ideas was to try setting up a homeserver on minimal extra costs, as long as it does not function as wished. Since this is my first project of this kind, I didnt want to dump more money than necessary. If I get it to work, then I would consider invest some money in bigger and more modern harddrives.

 

So: Is there a way to tell, if  the eclosures or the docking station support UASP?
I mean, by far the oldest drive, the Seagate 500GB, I bought somewhere around 2008 I guess, goes to sleep as intendet.


Second:

I would follow you about USB beeing the main problem. But since the internal M2.SSD hat the exact same problem when it was Disk 1 before, ist makes me believe, that there is another problem.

 

Is there no way to see, which data is read every couple of seconds, so the disk spins up again and again?

And: Is there a way to switch the both array disks? Making Disk 2 the new Disk 1 and the other way around?
If then, Disk 2 doesnt spin down, and Disk 1 does, it's proven that it is a hardware problem.
 

Does it happen if you disable Docker and VM Manager in Settings?

I stopped all dockers, then it still occurs. Didnt touch anything around VM ever, yet.

I will try to disable both in settings...



Aaaaand: Yep. Both deactivated via settings. Disk behaves like before, nothing changed.

1 minute ago, Carmino said:

Disk behaves like before, nothing changed

Do you have any evidence the disk is being read when Docker and VM Manager are disabled and nothing else is accessing your server? Do you have any mapped drives to your server anywhere on your network?

Now I set the statistic of the drives back to zero and ordered the drive to spin down manually, by clicking on the green indicator.
Now the disk is down and stays down. Maybe I tried to early before.

Parity disk does not react on the order to spin down. Maybe the drive or the docking station does not support that (?)

And I found this in the drive-protocol. Didn't knew there was one before.
Is that helpful?

 

I'm not sure what a "mapped drive" is. I have one shared folder on the disks, that I can access via Windows as a network-folder. But thats something else, I guess?

 

Sry I might appear pretty stupid. I'm just really new and unexperienced in this field.

 

Jan  5 14:23:34 UNSC-Gatekeeper emhttpd: spinning down /dev/sda
Jan  5 14:23:46 UNSC-Gatekeeper emhttpd: read SMART /dev/sda
Jan  5 18:38:54 UNSC-Gatekeeper emhttpd: spinning down /dev/sda
Jan  5 18:39:08 UNSC-Gatekeeper emhttpd: read SMART /dev/sda
Jan  5 18:45:11 UNSC-Gatekeeper emhttpd: spinning down /dev/sda

 

And now, after activating dockers, vm again, all my installed dockers are gone...

 

Even after waking up the drives manually.

The appdata is on the cache drives only...

7 minutes ago, Carmino said:

The appdata is on the cache drives only...

Is your cache drive working OK? 

Where is the docker.img file?   It should be on the cache pool in the 'system' share.   If it is missing and/or damaged then you are going to need to:

Recreate the docker.img file

Reinstall the docker containers with their settings intact

  

As far as I can see, the cache is working fine.

I can access the data and find the docker.img under cache/system/docker

How do I know if its damaged? It says size is 21.5GB, last changed ~20minutes ago

5 minutes ago, Carmino said:

As far as I can see, the cache is working fine.

I can access the data and find the docker.img under cache/system/docker

How do I know if its damaged? It says size is 21.5GB, last changed ~20minutes ago

 Is it possible that somewhere along the line you have managed to recreate an empty docker.img file?  If so then your container binaries are missing even though you might have the relevant appdata working set files.   If it checks out fine then you can just follow the steps to reinstall any containers that are missing with previous settings intact.

I wouldnt know how I should have done this...

I just got to settings, Docker and set the line "Enable docker" to "No" and hit apply. Same for the VMs.

 

If I look into the App-Store, under "installed apps" everything is marked as "insatlled". Not only under previous apps... (?)

 

 

EDIT: I was wrong: Not everything is displayed as installed. Only the plugins I guess...

I got a plugin called AppdataBackup. I try to restore it with that...

Edited January 5, 20242 yr by Carmino

7 minutes ago, Carmino said:

I got a plugin called AppdataBackup. I try to restore it with that...

That plugin does not reinstall docker container binaries - it just reinstalls the working set files for use with previously used containers.   The binaries are re-installed via Apps->Previous Apps->Docker.

You are totally right. Found that out as well and I'm on it. re-installing the dockers as you mentioned.

So...

 

Dockers are back and working after also renewing the custom network.

 

 

BUT:

 

The problem of a disk spinning up again and again has now switched from Disk 1 to Disk 2...

Unclear what all might have been changed since diagnostics was first posted several hours ago.

4 minutes ago, Carmino said:

The problem of a disk spinning up again and again has now switched from Disk 1 to Disk 2...

I suspect this is because a container is regularly accessing the drive.   You could try stopping individual containers to see if that changes behavior.

Couple of minutes later, I tried again to put both disks to sleep and voilá!

 

Both stay down now....

 

So, initital problem is solved, it seems, but I find it hard to describe, what the solution was.

De-activating Dockers and VMs, re-activating and then re-installing all dockers?

Nah... nevermind. Seems to have been a rare instance that it stayed down for a couple of minutes.

 

Now, as mentioned before, Disk 2 comes up again by itself.

 

@itimpi  By "stopping individial containers" you mean stopping the individual dockers, right?

Edited January 5, 20242 yr by Carmino

2 minutes ago, Carmino said:

@itimpi  By "stopping individial containers" you mean stopping the individual dockers, right?

Yes.

 

i would think the most likely candidate is Nextcloud as that has files on disk2, but that is just a guess as it also has files on other drives.