Update from 6.9 to 6.11.5 and got permission denied

[gandalf15]

Hello everyone

 

I run Unraid now for something like 2 years and all worked perfectly. I hardly do updates and I saw that I am "2 versions" behind and updated from 6.9 to 6.11.5.

Now my dockers have problems:
Syncthing cant sync anymore - when it tries to download files it gets permission denied.
 

I searched the web and it said "use chmod 777 yourfolder/"

I did that for the folders where syncthing downloads its files.

Now Radarr, Sonarr, Lydarr all cant import the files - also permission denied.

It looks like people have those problems since 6.10.

Someone can guide me how to fix that problem for all my dockers and folders?

I read something about the "New Permissions" under Tools, which I run currently on all disks and all User Shares.

Is this the way to go?

Thank you for your help!

[gandalf15]

I downgraded back to 6.9.2  and all works again.

I dont know but something is brocken with permissions on later releases.

Edited December 1, 2022 by gandalf15

[JorgeB]

Some good info here:

 

https://forums.unraid.net/topic/123901-plex-issues-upon-upgrade-to-6101/?do=findComment&comment=1138715

 

more specifically about rclone and sonar/radar/plex

https://forums.unraid.net/topic/75436-guide-how-to-use-rclone-to-mount-cloud-drives-and-play-files/page/115/#comment-1140504

[gandalf15]

Thanks I read those threads already. Sadly they dont help.

I am a bit disapointed, that an update causes that much of problems and there is no official FAQ how to solve that problem.

 

There are a lot of posts with the same problem. I hope there will be a official solution - without becoming a linux pro. If I was looking for a pro solution, I could go with a baremetal linux distro. I got unraid since it was user friendly. I hope it will move back into that direction.

 

Except my storj docker - every docker stopped working after the update.

 

Edited December 1, 2022 by gandalf15

[gandalf15]

On 12/1/2022 at 8:00 PM, gandalf15 said:

Thanks I read those threads already. Sadly they dont help.

I am a bit disapointed, that an update causes that much of problems and there is no official FAQ how to solve that problem.

 

There are a lot of posts with the same problem. I hope there will be a official solution - without becoming a linux pro. If I was looking for a pro solution, I could go with a baremetal linux distro. I got unraid since it was user friendly. I hope it will move back into that direction.

 

Except my storj docker - every docker stopped working after the update.

 

Is there a solution that works now or do I really have to make a clean install of everything?

[tawehruhle]

On 12/27/2022 at 4:06 PM, gandalf15 said:

Is there a solution that works now or do I really have to make a clean install of everything?

I had a similar problem after upgrading from 6.9 to 6.11 and I solved it by going to > Tools > "Docker Safe New Perms" and running. Everything worked after that completed.

[gandalf15]

On 1/3/2023 at 8:24 PM, tawehruhle said:

I had a similar problem after upgrading from 6.9 to 6.11 and I solved it by going to > Tools > "Docker Safe New Perms" and running. Everything worked after that completed.

how long did it take for you to complete and how much space?
I have 10 TB (a lot of tiny files) and I am scared if that could take days to complete?

 

Edit: Also I read that you should not run it on appsdata folder, did you include everything?

 

Edited January 5, 2023 by gandalf15

[tawehruhle]

On 1/5/2023 at 4:25 PM, gandalf15 said:

how long did it take for you to complete and how much space?
I have 10 TB (a lot of tiny files) and I am scared if that could take days to complete?

 

Edit: Also I read that you should not run it on appsdata folder, did you include everything?

 

I have 16TB and it took about 6-8mins I think.

Also if you look at the bootom of the page the appdata is excluded by default

It fixed most thing but my docker containers using appdata still has problems

I plan to try it on appdata (apparently not recommended) when I get home after a backupof apss ofcourse

[gandalf15]

8 hours ago, tawehruhle said:

I have 16TB and it took about 6-8mins I think.

Also if you look at the bootom of the page the appdata is excluded by default

It fixed most thing but my docker containers using appdata still has problems

I plan to try it on appdata (apparently not recommended) when I get home after a backupof apss ofcourse

Thank you very much for your testing. That is the problem I see and why I dont think its the solution: Dockers still dont work after doing that fix. I have some dockers running that I cant lose (Storj for an example, the node is 24 months old...)

[gandalf15]

@tawehruhle Do you have some news?

I simply dont have the capacity yet for a backup

[Redington]

did you get any benefit for upgrading from 6.9.2 to 6.11.5, well any benefit that would make you upgrade. I'm looking to do the same. going from 6.9.2 to 6.11.5.   I was a little surprised that I made a similar post & didn't get any help from the Unraid moderators. 

 

Any handy hint that are not already on this thread.  

 

HELP Upgrading OS from 6.9.2 to 6.11.5 best practice - General Support - Unraid

[JoeUnraidUser]

To fix the file ownership and permissions for my Dockers I do the following:

chown -cR nobody:users /mnt/user/appdata
chmod -cR ug+rwX,o-rwx /mnt/user/appdata

 

I set the following settings in each of my dockers to fix ownership and permission problems:

PUID = 99
PGID = 100
UMASK = 000

PUID of 99 equates to "nobody", PGID of 100 equates to "users", and UMASK of 000 allows for full access.

 

To fix the ownership and permissions of my array I do the following:

chown -cR nobody:users /mnt/disk[0-9]*
chmod -cR ug+rwX,o-rwx /mnt/disk[0-9]*

 

To fix the file ownership and permissions on my entire server I do the following:

chown -cR nobody:users /mnt/user/appdata /mnt/disk[0-9]*
chmod -cR ug+rwX,o-rwx /mnt/user/appdata /mnt/disk[0-9]*

 

You do not have stop your Dockers during the process of fixing the ownership and permissions of the array or the entire server.
 

Edited February 7, 2023 by JoeUnraidUser

[gandalf15]

3 hours ago, JoeUnraidUser said:

I run the following to fix the file ownership and permissions on my entire server.  It is safer than running "New Permissions" since it removes permissions for "other" which is more secure and keeps Dockers from failing.  The "Docker Safe New Perms" is sort of pointless since it does about the same thing as "New Permissions" but just leaves out appdata.

chown -cR nobody:users /mnt/user/appdata/* /mnt/disk[0-9]*
chmod -cR ug+rw,ug+X,o-rwx /mnt/user/appdata/* /mnt/disk[0-9]*

I'm sorry I did not notice your concerns in the past, I would have given you this information had I realized you had questions.

 

Update:

I hope you get my posts @limetech I am only trying to help you out.

 

Thank you very much. I actually found help on Reddit.

 

What I did was update, stopped all dockers and shut down docker service. I started then the docker safe new perms plugin. After 20 hours I canceld it and excluded 1 share (beside appdata), since it took too long and the docker had to get back online (glad it works).

 

After that I  run "sudo chown -Rc nobody:users /mnt/user/appdata"

 

When this was finished, I added to all docker the value PUID 99 and PGID 100 (some had it already).

 

Then my rclone mount made problems, since it was mounted as root:root. I then added PUID 99, PGID 100 AND UNMASK 000 to my mount script as an argument.

In the end it seems like all is working again.

 

To be honest, I have no idea if this was secure or/and right. But it works again, this is what counts (now) for me.

[JoeUnraidUser]

On 1/26/2023 at 5:44 AM, gandalf15 said:

 

Thank you very much. I actually found help on Reddit.

 

What I did was update, stopped all dockers and shut down docker service. I started then the docker safe new perms plugin. After 20 hours I canceld it and excluded 1 share (beside appdata), since it took too long and the docker had to get back online (glad it works).

 

After that I  run "sudo chown -Rc nobody:users /mnt/user/appdata"

 

When this was finished, I added to all docker the value PUID 99 and PGID 100 (some had it already).

 

Then my rclone mount made problems, since it was mounted as root:root. I then added PUID 99, PGID 100 AND UNMASK 000 to my mount script as an argument.

In the end it seems like all is working again.

 

To be honest, I have no idea if this was secure or/and right. But it works again, this is what counts (now) for me.

Glad you got it fixed.  I know some other people had some problems with PUID and PGID being set to 0 which is "root".  A PUID of 99 is "nobody" and PGID of 100 is "users" which is the correct way to do it.  I also like to set UMASK to 000.  If you have any permission problems in the future and need to run the commands to fix them, you don't need to stop your dockers.  Just let the commands run and use your dockers as normal.

 

 

Edited August 13, 2023 by JoeUnraidUser

[gandalf15]

7 hours ago, JoeUnraidUser said:

Glad you got it fixed.  I know some other people had some problems with PUID and PGID being set to 0 which is "root".  A PUID of 99 is "nobody" and PGID of 100 is "users" which is the correct way to do it.  I like to set UMASK to 007 because it leaves out access to "other" which makes it a more secure, but 000 works too.  If you have any permission problems in the future and need to run the commands to fix them, you don't need to stop your dockers.  Just let the commands run and use your dockers as normal.

 

 

That is a very good information - thank you.

I had to stop the fix for one share since it about 10 TB of very little files in the kb range - and that took 20 hours and wasnt finised. I might rerun it then just to make sure

[gandalf15]

@JorgeB

 

Could you please make JoeUnraidUser's post the solution. Seems better than mine (although I dont know if it works since I used the way I posted).

Thank you!

[JorgeB]

1 minute ago, gandalf15 said:

Could you please make JoeUnraidUser's post the solution.

Done

[gandalf15]

4 minutes ago, JorgeB said:

Done

Thank you!

 

[ados]

Went from 6.9 to 6.11.5 and had issues with that smaller Unraid system, particularly docker, rolled back and everything perfect.

Had to upgrade main big system because the license upgrade function wouldn't work and I needed pro, its basically broken everything.

Dockers don't work in assorted ways and I've spent a few hours reading forum suggestions which haven't worked.

I set no custom docker permissions, everything was template based.

The suggested app data permissions reset has broken more dockers, mainly Nextcloud, OpenProject and anything SQL.

Will try rolling back to 6.9 and stay there 😫

[Spyderturbo007]

On 1/26/2023 at 1:55 AM, JoeUnraidUser said:

To fix the file ownership and permissions for my Dockers I do the following:

chown -cR nobody:users /mnt/user/appdata
chmod -cR ug+rwX,o-rwx /mnt/user/appdata

 

I set the following settings in each of my dockers to fix ownership and permission problems:

PUID = 99
PGID = 100
UMASK = 000

PUID of 99 equates to "nobody", PGID of 100 equates to "users", and UMASK of 000 allows for full access.

 

To fix the ownership and permissions of my array I do the following:

chown -cR nobody:users /mnt/disk[0-9]*
chmod -cR ug+rwX,o-rwx /mnt/disk[0-9]*

 

To fix the file ownership and permissions on my entire server I do the following:

chown -cR nobody:users /mnt/user/appdata /mnt/disk[0-9]*
chmod -cR ug+rwX,o-rwx /mnt/user/appdata /mnt/disk[0-9]*

 

You do not have stop your Dockers during the process of fixing the ownership and permissions of the array or the entire server.
 

I was having the issue where some containers were unable to write to the array.  Someone told me to run the first set of commands and everything is all messed up now.  I am unable to open some of my docker container folders inside of windows.  For example, I get a permissions issues when trying to navigate to the AppData\EmbyServer folder.   Emby is also throwing errors trying to write its own configuration file backup, saying that the EmbyServer folder doesn't exist.

 

Is there a way to fix my issues?  This entire 6.9 -> 6.11 thing royally screwed up my server.  I wish the developers would have warned people about these issues or at least helped the people that are having these issues.

 

I've created threads in docker container forums, the individual program forums (Emby, Sonarr, etc) and multiple threads here with no resolution.  I don't know what to do other than nuking all the docker containers having issues and starting over again.

 

At this point I'm literally begging for help.  

[itimpi]

I suspect you may well have to redo your docker containers.   Many of them have specific requirements around permissions and running the commands you did against the appdata share would have messed these up.

[Spyderturbo007]

46 minutes ago, itimpi said:

I suspect you may well have to redo your docker containers.   Many of them have specific requirements around permissions and running the commands you did against the appdata share would have messed these up.

Is there a way to restore the permissions to how they were before and then fix them individually?  It appears to be Emby and Sonarr that were broken.  I have backups of the AppData folder.

[mrzablah]

On 1/26/2023 at 12:55 AM, JoeUnraidUser said:

To fix the file ownership and permissions for my Dockers I do the following:

chown -cR nobody:users /mnt/user/appdata
chmod -cR ug+rwX,o-rwx /mnt/user/appdata

 

I set the following settings in each of my dockers to fix ownership and permission problems:

PUID = 99
PGID = 100
UMASK = 000

PUID of 99 equates to "nobody", PGID of 100 equates to "users", and UMASK of 000 allows for full access.

 

To fix the ownership and permissions of my array I do the following:

chown -cR nobody:users /mnt/disk[0-9]*
chmod -cR ug+rwX,o-rwx /mnt/disk[0-9]*

 

To fix the file ownership and permissions on my entire server I do the following:

chown -cR nobody:users /mnt/user/appdata /mnt/disk[0-9]*
chmod -cR ug+rwX,o-rwx /mnt/user/appdata /mnt/disk[0-9]*

 

You do not have stop your Dockers during the process of fixing the ownership and permissions of the array or the entire server.
 

Thanks a lot! this totally fix me issue with permissions!!