Some questions regarding my Unraid-Setup


Recommended Posts

1

When I'm tying to write (and sometimes red) to/from the array, my CPU (E3-1231 v3) sits at 100% at all 8 threads.

The Copying process is fairly slow and this also makes all of my dockers temporarily unavailable.

 

I think this problem may partially be caused by my cache disk configuration.

 

2

I have 2 cache-drives installed.

1 ssd & 1 hdd. The ssd was installed a few days ago with a lot of red/write in the meantime. But it says there that writes are 0. Since all my dockers are on the HDD could this cause a problem?

And how should I configure this, that unraid uses the ssd as a primary cache-disk and not the hdd.

And will this solve my CPU-Usage problem?

 

3

I thougt about removing the hdd from the array and mount it with unassigned devices.

Can I then run the dockers from there & will they still be detected. Will I have to reinstall them?

If i will have to reinstall, can I save the config files from my (linuxserver) plex docker?

 

//sorry for spelling/grammar mistakes

Thanks!

Link to comment

Your cache description isn't entirely clear, but in any case, however you have it configured, it is a pool with no way to specify the disks separately when accessing them. Most people would never consider mixing hdd and ssd in the cache pool.

 

By "removing hdd from the array" I assume you actually mean removing it from the cache pool, not the parity array. People often use an Unassigned Device for dockers, etc. but an SSD is going to perform better with apps.

 

I really don't have enough details about your system to make specific recommendations about how to proceed, and getting your configuration changed will probably take some detailed work, but it can be done.

 

To give us a more complete idea of what you have and what might be happening, go to Tools - Diagnostics and attach the complete diagnostics zip file to your next post.

  • Like 1
Link to comment
6 minutes ago, johnnie.black said:

SSD was dropped from cache at some point in the past, likely dropped offline, you can re-add it but IMO not much point in running an hybrid SSD/HHD cache, I would use one or the other, depending on if you need speed or capacity.

 

So, i could just

stop the array, remove my hdd from cache, add it to unassigned devices, start array.

 

Then i'd have a ssd cache pool that i can upgrade with another ssd when i need to and all my dockers on the hdd.

Will unraid still recognize those dockers?

Link to comment
14 minutes ago, johnnie.black said:

Not currently since the SSD was removed from the pool, you can re-add it then remove the HDD, best to clear the SSD first though.

The SSD currently is in the pool.

It's the replacement for the old ssd that is completly broken and is not recognized any more and unraid won't start the array with it installed.

And on the new device are no files, as there are no writes listed. 
Or did I get something wrong?

Link to comment
1 hour ago, Einzigstes said:

The SSD currently is in the pool.

It's the replacement for the old ssd that is completly broken and is not recognized any more and unraid won't start the array with it installed.

And on the new device are no files, as there are no writes listed. 
Or did I get something wrong?

Can you currently read files from cache? If not then something is going to need fixing before you can have any hope of getting that data to an Unassigned Device. And just taking that HDD and moving it to Unassigned Devices isn't going to result in the pool data you want to be on Unassigned anyway. So as I said, some detailed work to do to reconfigure things.

 

I question whether moving the HDD to Unassigned for your dockers is the right approach if that is what you had in mind. I would leave the HDD as cache and put the SSD as Unassigned for apps if that is the approach you want to go for.

 

If you don't mind spending some money I would just put that HDD in the parity array and get 1 or 2 larger SSDs for the cache pool and forget about running apps on an Unassigned Device, since it will require some reconfiguration of how you use docker and install containers.

 

  • Like 1
Link to comment
1 minute ago, trurl said:

Can you currently read files from cache? If not then something is going to need fixing before you can have any hope of getting that data to an Unassigned Device. And just taking that HDD and moving it to Unassigned Devices isn't going to result in the pool data you want to be on Unassigned anyway. So as I said, some detailed work to do to reconfigure things.

 

I question whether moving the HDD to Unassigned for your dockers is the right approach if that is what you had in mind. I would leave the HDD as cache and put the SSD as Unassigned for apps if that is the approach you want to go for.

 

If you don't mind spending some money I would just put that HDD in the parity array and get 2 large SSDs for the cache pool and forget about running apps on an Unassigned Device, since it will require some reconfiguration of how you use docker and install containers.

 

As my dockers are still working properly (when there is no file beeing written to the server) I assume Reading from cache is possible, yes!

If unraid really writes files to the cache, I don't know... Because of the CPU-usage i have some doubts there.

 

I guess it doesn't really matter which disk is for the cache pool & which for the unassigned device.

 

I could of course buy a second ssd with 240gb or something and remove the HDD from the pool. Wouldn't really be a problem.

But what I dont know is if the CPU Usage will be lower on writing. 
It should just write to the cache disk and the cpu should technically not be needed that much?

 

Or could my limitation now be the read/write capacity of my Hdd and the dockers' read/write AND my copy process is too much for it to handle?

But wouldn't then my copy-process just slow way down and instead of the cpu working heavily?

 

 

-----

 

I just can't understand why this would ever happen and I have some doubt if the cause of this isn't somerwhere else...

 

Thank you!

Link to comment
3 hours ago, Einzigstes said:

The SSD currently is in the pool.

It's not, despite still appearing on the main page, it was deleted from the pool, that's why there are no writes to it:

 

Jan 12 16:34:24 Server1 kernel: BTRFS warning (device sdc1): devid 3 uuid 059b7e98-efae-4354-80b1-d964777a4211 is missing
Jan 12 16:34:24 Server1 kernel: BTRFS info (device sdc1): bdev (null) errs: wr 2541, rd 1, flush 0, corrupt 0, gen 0
Jan 12 16:34:27 Server1 emhttpd: shcmd (49): /sbin/btrfs filesystem resize 2:max /mnt/cache
Jan 12 16:34:27 Server1 root: Resize '/mnt/cache' of '2:max'
Jan 12 16:34:27 Server1 emhttpd: shcmd (50): /sbin/btrfs device delete missing /mnt/cache && /sbin/btrfs balance start /mnt/cache &

 

Link to comment

Your mounted capacity for cache doesn't really make any sense for either or both of those 2 disks so something not right.

 

I don't know if your dockers are working or not. And the diagnostics from your old version don't help me figure out which disks they would be using anyway. I know your system share is configured to move files off cache so the docker image may not even be on cache. And your appdata share is configured to write new files to cache and not move them but that doesn't mean some of it wasn't already on the array.

 

Whether your storage configuration problems are the cause of your other problems is unclear.

 

Go to Shares - User Shares and click on Compute All at the bottom of the page, then wait for it to get the results and post a screenshot.

  • Like 1
Link to comment
1 minute ago, johnnie.black said:

It's not, despite still appearing on the main page, it was deleted from the pool, that's why there are no writes to it:

 


Jan 12 16:34:24 Server1 kernel: BTRFS warning (device sdc1): devid 3 uuid 059b7e98-efae-4354-80b1-d964777a4211 is missing
Jan 12 16:34:24 Server1 kernel: BTRFS info (device sdc1): bdev (null) errs: wr 2541, rd 1, flush 0, corrupt 0, gen 0
Jan 12 16:34:27 Server1 emhttpd: shcmd (49): /sbin/btrfs filesystem resize 2:max /mnt/cache
Jan 12 16:34:27 Server1 root: Resize '/mnt/cache' of '2:max'
Jan 12 16:34:27 Server1 emhttpd: shcmd (50): /sbin/btrfs device delete missing /mnt/cache && /sbin/btrfs balance start /mnt/cache &

 

Well that's weird...

 

So, what can I do now?
Readding the broken disk? The array wouldn't start with it, that's why I removed it in the first place.

Link to comment
1 hour ago, trurl said:

Your mounted capacity for cache doesn't really make any sense for either or both of those 2 disks so something not right.

 

I don't know if your dockers are working or not. And the diagnostics from your old version don't help me figure out which disks they would be using anyway. I know your system share is configured to move files off cache so the docker image may not even be on cache. And your appdata share is configured to write new files to cache and not move them but that doesn't mean some of it wasn't already on the array.

 

Whether your storage configuration problems are the cause of your other problems is unclear.

 

Go to Shares - User Shares and click on Compute All at the bottom of the page, then wait for it to get the results and post a screenshot.

 

Here's the Screenshot!

I just had a look into my server and it could be a sata-power cable came loose in my cache-hdd with SATA still connected.

 

Could it be that Unraid was detecting the hdd but it didn't spin up and unraid started writing and reading directly to/from the array?

I will have to test if it get's any better with the cpu-usage...

Even though I'd like to remove the HDD from the cachepool!

Link to comment
1 hour ago, johnnie.black said:

I would recommend you backup your cache, since it's in a state that can go unpredictable and unmountable, reformat with just the device you want and restore the data, you can use this for help with the backup/restore.

Ok, but if I just search for the 2 config files I need and download those & just remove the cache disks after I run the Mover....

This shouldn't have any consequences on the files in my array, right?
Would be a bit of work to get some of the dockers back up again, but i think I'd like that bit of a fresh start

Link to comment
Just now, trurl said:

Your screenshot shows appdata all on cache but system share has the same amount on cache and disk2, probably a duplicate of your docker image.

Maybe I missed it in a previous post but which 2 config files?

You didn't miss anything, just saying that i could get the config files that are important and everything else in the cache pool can be deleted/ the disks can be removed (If this wouldn't bring any more problems with it which i don't know about. I just assume that my files on the array are not affected by this?!)

Link to comment
Just now, trurl said:

Which files would those be?

-The Plex Configuration mabye

-MotionEye camera profiles

-Some stuff from the mineos-docker

 

That schould be all.. Nothing really important on these Drives!

Wanted to rebuild my Plex Library anyways.

Could back up the estimated 10GB of files at any time and remove the disks without any loss of important data

Link to comment

I now backed everything up and added the hdd to the array.

I ordered a 500GB ssd for the cache pool. When this arrives I will reinstall all my dockers to the cache pool only.
 

When Installing a new docker, will the "cache-only" setting in the appdata share be enough for the docker do just be in the cache pool or will I sometimes have to change paths manually before installing the docker?

Link to comment
2 hours ago, Einzigstes said:

When Installing a new docker, will the "cache-only" setting in the appdata share be enough for the docker do just be in the cache pool or will I sometimes have to change paths manually before installing the docker?

Yes, but I believe some dockers work best with /mnt/cache instead of /mnt/user, I always prefer using /mnt/cache for the paths.

Link to comment
3 hours ago, Einzigstes said:

I now backed everything up and added the hdd to the array.

I ordered a 500GB ssd for the cache pool. When this arrives I will reinstall all my dockers to the cache pool only.
 

When Installing a new docker, will the "cache-only" setting in the appdata share be enough for the docker do just be in the cache pool or will I sometimes have to change paths manually before installing the docker?

appdata is currently all on cache as seen above and since that share is set to cache-only, that is the only place new files will be written.

 

But system is set to cache-yes which means it will be moved to the array, and as also seen above there is likely a duplicate docker image on disk2. The system share should be set to cache-only also, your docker image deleted (from Settings - Docker), and from the other disk if it is still there. Then you can recreate your docker image and go to Apps - Previous Apps to add them back.

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.