3 hours ago, JorgeB said:

A few users are seeing a "Wrong Pool State, invalid expansion" error with a pool after upgrading, this usually means that the pool had an incorrect cfg file, but v6.12 didn't check or care about it, you can easily resolve the issue by just letting v7 re-import the pool, to do that:

 

• on main, click on that pool and then "Remove Pool"
• back on main, create a new pool with the same name and number of slots
• assign the pool device(s), leave the filesystem set to auto
• start the array to import the pool

 

 

Is this also a solution for my problem that after an update from 6.12.14 to stable 7.0.0 my my pool cache drive (where appdata, system and domains are stored) wasn't ready, and the system wanted it to be formatted?

As I understand it, it will wipe my whole cache drive, so it's there a way to fix that ?

Updated and have been running stable for 3 hours now. No errors  

37 minutes ago, rogales said:

wasn't ready, and the system wanted it to be formatted?

Not clear to me what error you are seeing, but if the pool is valid, doing that will just re-import the existing pool and data, and even if it doesn't work, it won't delete anything, but you can also start a new thread in the general support forum and post the diagnostics and the exact error you are experiencing.

2 minutes ago, JorgeB said:

Not clear to me what error you are seeing, but if the pool is valid, doing that will just re-import the existing pool and data, and even if it doesn't work, it won't delete anything, but you can also start a new thread in the general support forum and post the diagnostics and the exact error you are experiencing.

Thank You for Your reply.
Will try one more time, but now I will save screenshot before, and after the update so it would be more clear for You to understand if this error occur.

For what it's worth: Just updated to 7.0.0 with no issues.

 

One small remark, I had one docker container NOT set to auto-start, but it was set to auto-start after the update (like all the others).

...I am like 99% sure it was not set to auto-start .

 

Or in other words: probably it was, but perhaps something to monitor for others..

updated. booted and my cache pool threw the 'invalid pool state' error. Mild panic, read the forums, sorted it. 

 

Click the pool name. Remove it. Recreate it with the same name. Add the drives as per last pool. Done - been stable since. 

Edited January 10, 20251 yr by Adamau5

@JorgeB

4 hours ago, mraerosmith said:

Went up to 7.0 without a hitch, nice

Feature requests are handled in the dedicated subforum : https://forums.unraid.net/forum/53-feature-requests/

There are already several topics there and arguments why it is not a priority.

Upgraded from 7.0.0 RC, so far, no issues, good job everyone.

Upgraded from 6.12.14 without any issues!

 

Thank you for your dedicated and ongoing efforts

 

//Thomas

I see the notes that if you're using data-dir for docker images and wanting to use a zfs pool as the target storage location that we should select overlay2. In the docker settings, I see that 'native' is one option and shows a zfs option for zfs-formatted file systems. Is there a reason overlay2 is better than the native/zfs option for zfs pools? I am not currently using a pool for my appdata and looking to gain some mirror data protection with a zpool. Thank you.

Updated without problems 👌

Updated 2 machines, smooth. Fixed the iGPU kernel problem as well. NICE

 

I guess back to sleep now

Edited January 10, 20251 yr by DivideBy0

Upgraded last night but the mover isn't working. Seemed to be a common issue with the beta with the mover tuner plugin but I am not using that plugin. The log shows: 

 

shcmd (45584): /usr/local/sbin/mover start &> /dev/null &

13 minutes ago, dopeytree said:

Pretty minor thing but noticed download not working in file manager. It gives a little false file that's a few kbs instead of a 24MB mp4 file. 

What 'file manager'--- Dynamix or Windows or some other file manager? 

 

And what was the destination of this download?

25 minutes ago, tojoski said:

Upgraded last night but the mover isn't working. Seemed to be a common issue with the beta with the mover tuner plugin but I am not using that plugin. The log shows: 

 

shcmd (45584): /usr/local/sbin/mover start &> /dev/null &

Please create a post in the general support forum, enable the mover loggin, run the mover, and post the diagnostics and the name of the share(s) you expect moved.

I have a ZFS pool for my cache drives (not the standard BTRFS), is this supported from an upgrade from 6.12.14 -> 7.0 ? Thanks

Upgraded from 6.12.14.  No issues.  Waiting to "upgrade" zfs pools until I am certain everything is stable.  No problems with VM's or Dockers.

Upgraded backup system without issue from 6.12.14 - intel i3-6100, nothing fancy here. Did have to uninstall/reinstall the intel-gpu-top plugin.

 

Main system upgrade from 6.12.14 so far is failing to boot and looking like something hosed on the flash drive. This is an AMD 5600x on x570-d4u - initially looks like it's failing to mount /dev/sda1 and it gets stuck here. Starting investigation and will open a separate thread if I can't figure this out.

 

update 1: There are two other USB devices attached - the UPS (Cyberpower) , and a Rainforest EMU-2. Disconnecting both devices allowed the system to boot without issue, though oddly the HomeAssistant docker did not start (and had no unusual logs). Manually started just fine. Plugged both USB devices back in - the Rainforest is working properly. NUT plugin is not starting yet, but that's probably because the UPS wasn't there at boot so chasing that now.

 

Still very confused by why this happened, but at least it was relatively straight-forward to troubleshoot and get to a better place.

 

update 2: Rebooted again with both USB devices attached. This time, it automatically rebooted somewhere in the startup process (I was distracted and did not see it happen, only noticed when it went back to POST) but then that second boot worked through cleanly and the system is back up fully, NUT sees the battery, HomeAssistant started up automatically, and everything seems back to working properly. Meanwhile, this has reinforced the value of IPMI for me...

 

update 3: It turns out usb 4-2 is in fact the flash drive. I did not change ports. I did, however, unplug the flash drive and attach it to a Windows computer to check it wasn't showing errors there or anything, and I did this before the successful boot without the other two USB devices. It took an unusual amount of time to unmount. Makes me wonder if Windows quietly did something. It DID successfully start to boot,  Unraid though, so really no idea at this point. Very weird.

Edited January 10, 20251 yr by _cjd_
adding progress notes

10 hours ago, zoggy said:

a 2G prob works with a clean install (no plugins/no baggage of logs-notifications-backups).. but yeah prob doesnt hurt to bump min requirement to 4G these days as everything keeps growing....

 

 

but anyways to note, upgraded to 7.0 from 7.0rc2 without any issues. using 4G usb stick

I'm still on original 2GB myself. Any recommended options if I have to buy one?

 

nvm: Just saw this

Edited January 10, 20251 yr by betaman

Update was working fine from 6.12.14 to 7.0.0. on a Qnap NAS

But it needed one extra boot to start and one more extra boot for the nut-addon to work again.

 

Now I could remove my dummy usb-drive to have the array working. 👍

 

The only thing what is not working is the "stats page". I deleted cookies and browser cache, but this didn't help. Could it be that stats is not working because of the "none" array drives?

2 hours ago, chaosclarity said:

I have a ZFS pool for my cache drives (not the standard BTRFS), is this supported from an upgrade from 6.12.14 -> 7.0 ? Thanks

Yes, zfs was already supported in v6.12, and it still is.

I literally just upgraded to Unraid 7.0 from 6.12.14 and it took about 5 minutes with a restart. After the restart, my server came back up without any errors. I checked all my docker containers to make sure they started up properly and were accessible. All of my containers were fine. All of my plugins with the exception of NerdTools were fine. I received a update notification from the community app store for NerdTools and that update notification was to actually uninstall NerdTools. its deprecated in Unraid 7, but that solution was easy. There's another plugin in the community store called 'Python3' and it works as a nice replacement for NerdTools and its compatible with Unraid 7. In case you guys are interested, I run 14 docker containers that include the following:

Immich
Overseerr

Plex Pass

Qbittorent

Sonarr

Radarr

Prowlarr

Gluetun

flaresloverr

Postgres Backup

MySpeed

Krusader

 

All containers started up without any issues and are fully accessible. My array is also up and running without any issues. My NVMe cache drive appears to fine and also my NVMe appdata drive appears to be fine as well. I'm actually pretty impressed with how smooth and quick the upgrade went. I was a little nervous to upgrade after reading a few horror stories from others about past upgrades, but so far, the success stories about upgrading to 7.0 seem to be true. It was absolutely painless.

 

I just updated to Unraid OS version 7.0.0 from version 6.12.14. After a reboot the array is stopped and when I try to start the array I keep getting this message: 

Wrong Pool State

cache - too many wrong or missing devices

I had to downgrade to version 6.12.14 twice after failed updates. 

Is there any work around to get the array started normally like in version 6.12.14?

Any help will be appreciated. 

Thanks. 

 

 

Unraid error.docx

Edited January 10, 20251 yr by deshpremi

I recently attempted to upgrade my Unraid server from version 6.12.14 to 7.0.0 and encountered several critical issues that forced me to revert back to 6.12.14. I’d like to share my findings in case others experience similar problems or can offer insights.

 

System Configuration:
Motherboard: ASUS Pro WS W680-ACE IPMI (includes an onboard IPMI card)
HBA: Broadcom HBA 9500i
MB NIC: Intel(R) Ethernet Controller I226-LM

 

After the upgrade, all 16 drives connected to the Broadcom HBA 9500i were no longer detected.
I suspect this may be due to missing or incompatible drivers for the HBA in Unraid 7.0.0.

 

Network Misconfiguration:
On the second reboot after upgrading, the primary network interface (eth0) was incorrectly assigned to a virtual USB NIC (CDC Ethernet Device) or possibly to the IPMI card instead of the Intel Ethernet Controller I226-LM. As a result, the server was unreachable over the network, and I could not access it remotely.

I found a similar issue discussed in a forum thread titled "No Network After Updating: eth0 Assigned to Virtual USB NIC", where the solution involves deleting the network.cfg and network-rules.cfg files and letting Unraid regenerate them upon reboot. However, the issue persisted.

 

IPMI Card Impact:

The presence of an onboard IPMI card may have contributed to the network misassignment, as Unraid seemed to misinterpret the available interfaces during boot.

 

Power-Related Quirk:
A particularly puzzling issue I encountered was that after issuing the poweroff or reboot commands, the server wouldn't complete the shutdown or restart process properly. Instead of gracefully powering down or rebooting, I had to manually press the power or reset button to get the system to actually turn off or restart. This unusual behavior was something I hadn't experienced before, and it was particularly frustrating given that the system appeared to be responding to other commands, but the shutdown sequence seemed to stall.
 

After Reverting to 6.12.14:
Drives connected to the Broadcom HBA are detected.
Network interfaces are assigned correctly, with the Intel NIC functioning as expected.

Power-related quirk is gone

 

Additional Note:
Ironically, I upgraded a second server, which is my test server, to Unraid 7.0.0, and it went flawlessly without any issues. This suggests the problem might be specific to the configuration of the Pro WS W680-ACE IPMI or its hardware components.

 

Conclusion:
It seems that the new Unraid 7.0.0 release may have compatibility issues with:

Broadcom HBA 9500i drivers.

Implementing interface binding during upgrades could help ensure proper network interface prioritization, especially in systems with multiple NICs, virtual USB NICs, or IPMI cards.

 

I hope this issue will be addressed in a future patch to allow seamless upgrades for systems with similar configurations.
 

angel-diagnostics-20250110-0209.zip

Edited January 10, 20251 yr by Zerax
added power-related quirk