Upgraded Cache - btrfs Device Replace Did Not Start


Go to solution Solved by JorgeB,

Recommended Posts

I recently upgraded to v6.11.5 and installed an NVMe drive to replace the current SATA SSD cache drive using these instructions: 

 

  1. Stop the array - Did this.

  2. On the main page click on the cache device you want to replace/upgrade and select the new one from the drop down list (any data on the new device will be deleted) - Did this.

  3. Start the array - Did this, however, the array did not start immediately.  I received message from the browser that the page needed to be refreshed.  When I did so, the Main page showed the array to still be offline.  I opened the Main page in a different tab and still had the same result.  There was no indication that anything was taking place and the only buttons available at the bottom of the Main page were Reboot, Shutdown, and Sleep IIRC.  I rebooted and lost GUI access.  The login screen would occasionally appear but would provide a 503 Service Unavailable error.  This continued for 4 hours, during which I would check randomly to see if the status changed.  After ~4 hours, I was able to login and GUI access was restored.  The array was online without me turning it back on again.  All buttons were available at the bottom and the Format button was added.  The new NVMe drive was set to Cache but was unmountable because it had not been formatted.  I ran the formatting process.

  4. A btrfs device replace will begin, wait for cache activity to stop, the stop array button will be inhibited during the operation, this can take some time depending on how much data is on the pool and how fast your devices are. - This and the remaining steps did not take place.

 

Is there something else I was supposed to do in order to have my old appdata etc. move to the new cache drive?  If I ran the Kluth CA Backup/Restore Appdata prior to the upgrade, can I do the restore on the new drive?  Thanks in advance!

cylon-diagnostics-20230326-1339.zip

Link to comment
21 minutes ago, rockbox948 said:

I received message from the browser that the page needed to be refreshed.  When I did so, the Main page showed the array to still be offline

Urks! 😞

This is an evil  and known bug that happens with some browsers (mainly firefox).

The only correct answer to this is: NO! (cancel!)

 

If you agree, shit happens (as you have already noticed now the hard way).

 

 

Link to comment
12 minutes ago, MAM59 said:

Urks! 😞

This is an evil  and known bug that happens with some browsers (mainly firefox).

The only correct answer to this is: NO! (cancel!)

 

If you agree, shit happens (as you have already noticed now the hard way).

 

 

Well hell...  I was definitely using Firefox so that explains why things didn't go as they were supposed to.  What is the remedy?

Link to comment
1 minute ago, rockbox948 said:

"Next time"?

I am talking about the nagging window that popped up and asked you to resend the request.

THAT is a bug and should not happen. Cancel it!

Else you will cancel your last command (stop the array) and the gui will become "strange".

 

This has actually nothing to do with "erasing" the cache, it just canceled what you were just doing. If the popup appeared after the "erase" command, erase never happened.

 

So it would be better for you to start again and when the popup comes, cancel it and continue as described.

 

Link to comment

no, read the instructions carefully!

They do not apply to your situation, they are meant for people the are running TWO cache SSDs combined as a RAID Array.

There you can pull out one, put in a new one and the old data will be restored automatically.

 

But you only have ONE drive to be replaced.

You COULD add this new drive to the old pool, wait until it is synced, take out die old drive and degregate the pool back to a single drive.

But usually it is much easier just to stop all dockers and VMs and copy all data from the cache to a folder on the Array.

 

Anyway, you have switched already and i assume, the machine is back up again and working?

So no need to rewind something.

 

If you want to erase the old ssd, just reinsert it, use the plugin "unassigned devices" to "preclear" it (this will erase everything and wipe it out with zeros).

 

I dont understand your current situation and what you want to do now (standard excuse: "english is not my natural language")

 

Link to comment
6 hours ago, MAM59 said:

no, read the instructions carefully!

They do not apply to your situation, they are meant for people the are running TWO cache SSDs combined as a RAID Array.

There you can pull out one, put in a new one and the old data will be restored automatically.

 

But you only have ONE drive to be replaced.

You COULD add this new drive to the old pool, wait until it is synced, take out die old drive and degregate the pool back to a single drive.

But usually it is much easier just to stop all dockers and VMs and copy all data from the cache to a folder on the Array.

 

Anyway, you have switched already and i assume, the machine is back up again and working?

So no need to rewind something.

 

If you want to erase the old ssd, just reinsert it, use the plugin "unassigned devices" to "preclear" it (this will erase everything and wipe it out with zeros).

 

I dont understand your current situation and what you want to do now (standard excuse: "english is not my natural language")

 

You are correct, I completely missed the part about not working on a single device pool.

 

The system is back up but without the old appdata.

Edited by rockbox948
Link to comment

if i would be you, I would not panic yet. I guess, all data is still on the old Samsung 840 and is just waiting to be moved over to the new cache drive.

(Unless you have formatted the old one already)

 

* Stop all dockers and VMs (NOT the array!)

* Mount the old Disk with unassigned devices, use the file manager plug in (or use the shell with "mc") to copy over everything from the old drive. 

* restart Dockers and VMs (they should be fine again now)

Link to comment
  • Solution

Under normal circumstances the 2nd option would be easier, unless something goes wrong during the balance, so always good to make sure backups are up to date before starting, then add the NVMe device to the cache pool, wait for the balance to finish, it will finish when the cache activity stops and you can stop the array, stop array, unassign the 840 EVO SSD, start array, a new balance will start to delete the old device, once that finishes you are done.

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.