Removing disk from Array, adding a new one and more


piratx

Recommended Posts

Hello, I am sure this might have been answered in the past but searching for some minutes now and didn't find anything to match my setup.

 

So I have unRAID 6.9.2 with one Parity disks and 10 disks. Little note here, some 4tb disk in the past had issues so I removed it from my array. Cause of this I have now Parity Disk 1-7 and then 9-11, Disk 8 is missing. 

 

1658802855_ScreenShot2021-11-16at7_51_49AM.thumb.png.27ccf4746f0db1807302e34bc8ac3078.png

 

Now, I want to fix this for a while now but my day was full for the last months.. but I have a new 8tb disk to add in there.

What I have done "wrong" in the past was the addition of a NVMe SSD on my array (Disk 11), I've added that to keep my VMs and I couldn't do it with a second pool back then.

 

So I need to remove DIsk 11 from the array and fix Disk 8 which is missing.. problem here is my hardware won't let me add another disk so I need to remove an old one so I can do this.

 

That would be a disk which is giving me this error for some years now.

1265965812_ScreenShot2021-11-16at7_50_50AM.thumb.png.11febd3774ffba944988a46132c785b4.png

 

How should I do the above?

 

I have tried to remove stuff from the Disk 11 (nvme) via a crazy way.

I've excluded the disk from all the shares but I can still "see" files on it (domains, appdata).

 

An the last thing. Should I continue using Cache pool with 2 SSDs or I can do that with one only?

Depends of what I am using those SSDs for I assume?

 

Only : Cache for appdata / Yes : Cache for my main shares. So everything I am copying over the network to array it goes through the SSDs.

 

Sorry for the long text and I 'm looking forward for your replies!

:)

Edited by piratx
Link to comment
9 hours ago, piratx said:

unRAID 9.6.2 with one Parity disks and 10 disks. Little note here, some 4tb disk in the past had issues so I removed it from my array. Cause of this I have now Parity Disk 1-7 and then 9-11, Disk 8 is missing. 

6.9.2?

 

Your screenshot shows the array started with no disks missing. I think you must mean that disk8 is not assigned. A missing disk would be indicated by a red X and its contents would be emulated. But we can "fill that gap" if that is what you want.

 

How exactly did you remove that disk8? New Config with parity rebuild?

 

9 hours ago, piratx said:

a disk which is giving me this error for some years now.

Not clear which disk you are referring to here, but UDMA CRC Errors are connection problems, not disk problems. Probably nothing wrong with the disk, There are some good arguments to be made for not having an SSD in an array with HDDs though.

 

9 hours ago, piratx said:

tried to remove stuff from the Disk 11 (nvme) via a crazy way.

I've excluded the disk from all the shares but I can still "see" files on it (domains, appdata)

Except for the Use cache settings, which determines if and how files get moved from/to cache to/from array, the other User Share settings are about where new files get written and won't affect existing files.

 

Do you mean you want to make disk11 a pool instead of an array disk, or do you mean you don't want it in the server at all?

 

Attach Diagnostics to your NEXT post in this thread.

 

 

  • Thanks 1
Link to comment

Hello and thank you for your answer!

 

I typed the version wrong.. sshh sorry :)

 

I have removed the disk cause of problems and I placed a new 8tb instead of 4tb the previous one was. Never did "New config" so far. So no I just left the slot "open" I suppose!

Ah sorry Disk 5 is giving me the UDMA CRC errors (WDC_WD40EFRX-68N32N0_WD) I have some disks which I installed from external 8tb disks (all WD) and I got a cheat from a friend for them to work properly on my setup.

 

I want to keep nvme disk 11 on its own pool, probably for appdata (will that make things faster) and VMs..

Diagnostics included

eleven-diagnostics-20211116-1735.zip

Link to comment
14 minutes ago, piratx said:

Never did "New config" so far. So no I just left the slot "open" I suppose!

Not possible, disk8 would still be considered part of the array, and its contents would be emulated. Did you perhaps "start from scratch" with a new install and no disks assigned? Are you sure you have valid parity?

 

 

Link to comment

Nothing you have said makes any difference regarding removing a disk from the parity array. You have to rebuild parity so the removed disk is no longer part of the parity calculation. Doesn't matter at all what was on the disk if it was part of the parity array.

 

1 minute ago, trurl said:

How do you know?

Have you had a parity check since then that had exactly zero sync errors?

Link to comment

Why did you cancel those most recent parity checks?

Nov 16 07:11:06 eleven emhttpd: unclean shutdown detected

This by itself probably means your parity isn't completely valid.

 

15 minutes ago, trurl said:

Did you perhaps "start from scratch" with a new install and no disks assigned?

Or did you boot up at some point and discover that none of your disks were assigned and so you had to reassign them all?

Link to comment

Removing disk11 from the array is going to require a parity rebuild anyway so how you got to this point I guess is moot. And since you will be doing New Config and rebuilding parity you can assign the other disks however you want.

 

Go to User Shares, click the Compute All button, wait for the result, then post a screenshot.

 

 

 

Link to comment
7 hours ago, piratx said:

keep nvme disk 11 on its own pool, probably for appdata (will that make things faster) and VMs..

Ideally, your appdata, domains, system shares would be all on some pool and not in the parity array, so your dockers/VMs would perform better and not keep array disks spunup.

 

But it doesn't look like that nvme has the capacity for all of that. Why is your docker.img so large? Are you sure you don't have some application writing into it instead of to mapped storage?

16 hours ago, piratx said:

my hardware won't let me add another disk

Why? Seems like you don't really have enough capacity in fast storage to support everything you are trying to do with dockers/VMs, so you are going to have to make some compromises. I don't know what you are doing with all that so can't give detailed advice.

 

Does your domains really need to be that large? VMs can access Unraid storage so you don't really need to create vdisks for all of their files.

 

Assuming you only have user share paths and not disk paths for any of your dockers/VMs, it should be possible to put that nvme (disk11) in its own pool with its data intact since all its files could still be accessed with the same user share paths.

17 hours ago, piratx said:

Yes : Cache for my main shares. So everything I am copying over the network to array it goes through the SSDs.

One compromise you might consider is not caching some of your main shares, so you can use some of the capacity of those cache SSDs as fast storage for your other purposes. And/Or you might consider changing cache pool to raid0 so you get more space that way.

  • Thanks 1
Link to comment
7 hours ago, trurl said:

Why is your docker.img so large? Are you sure you don't have some application writing into it instead of to mapped storage?

On 11/16/2021 at 7:58 AM, piratx said:

I've increased docker.img cause it was giving me status reports that it's going to be full from time to time.. Something like this:

eleven: Alert [ELEVEN] - Docker image disk utilization of 92%
Docker utilization of image file /mnt/user/system/docker/docker.img

 

7 hours ago, trurl said:

Does your domains really need to be that large?

It's just tests! I've created one of those VMs (the bigger one 200GB) to make some docker setup tests.. so I won't mess up with my original setup.. 

So that can be moved to the array for future reference or deleted if I keep just my docker compose configs etc. It's not really an issue.

 

So, if it's a matter of NVMe capacity, I could easily swap that one with a 512GB.

 

So, if I got those correct, in that NMVe it will be my appdata and VMs in a new pool.

I am removing or not my 4TB disks (disk 5) with the UDMA CRC error and adding my new 8TB in there as well, can be placed on disk 8 which is missing (as a number from the list?)

After I do that I run New config so the changes pass to the array.. 

 

What's the best way to move my appdata into the NMVe and only have it there? I recall last time I checked my appdata is 42-45GB (Plex library uses almost ~40GB of that?) also, when I'll have appdata in only one disk, I will need backup, or I can use my plugin like before?

image.thumb.png.37bc3a3a067fab7124138e6ecb058939.png

 

Thanks for all the help, I've connected to discord channel as well :)

Link to comment

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.