Backup plugin start failing

[Barlog9511 17]

Hi 

The backup and restore plugin works fine but it started failing 2 days ago

The Emby server version is  4.8.10.0 and the plugin version is 1.6.7.0

backup folder is OK 

Thanks

embyserver.txt

[Luke 42077]

Quote

    SQLitePCL.pretty.SQLiteException: 8458: disk I/O error
 

Hi, is your disk full or hard drive having a problem?

[robsch 45]

I started to have issues as well.

 

 

 

Strangely it makes the folder, it just fails to put anything IN the folder.    The server it is writing to has 60tb free, and I the permissions are properly set. I have manually mapped the drive to that machine and I can write/delete files and directories into that folder.

[Barlog9511 17]

On 11/20/2024 at 2:17 AM, Luke said:

Hi, is your disk full or hard drive having a problem?

Hi @Luke
My HDD is OK nothing change , this plugin just stop working

[Neminem 1518]

@robschThis look fun, you are not keeping any backups

What happens if you press restore ?

Mine did a backup today, i'm kepping 30 days.

 

Edited November 21, 2024 by Neminem

[robsch 45]

7 minutes ago, Neminem said:

@robschThis look fun, you are not keeping any backups

What happens if you press restore ?

Mine did a backup today, i'm kepping 30 days.

 

I had it to 10, but I changed it to 0, which I thought might of been "Unlimited" I have no options to restore anything. I changed it to 10, restarted Emby, and still get the same exact error.

 

[Neminem 1518]

Did you run the Emby Server Backup task, after editing the amount of backups to keep ?

[Neminem 1518]

Also make sure that emby can see that network device.

I do local backups, and offload them to my backup nas, for more than 30 days of retention.

On my backup nas I have 6 month of rolling backups.

[Luke 42077]

6 hours ago, Barlog9511 said:

Hi @Luke
My HDD is OK nothing change , this plugin just stop working

Have you tried scanning your drive for potential hardware problems?

[Barlog9511 17]

1 minute ago, Luke said:

Have you tried scanning your drive for potential hardware problems?

in my case, my main video, backup, and metadata folder are NFS shares, video and metadata are fine and only the backup is causing the problem, 
our storage is checked regularly and shows no issues. I have also tried manually accessing the data and creating additional folders or some test data and everything works

[Luke 42077]

9 hours ago, Barlog9511 said:

in my case, my main video, backup, and metadata folder are NFS shares, video and metadata are fine and only the backup is causing the problem, 
our storage is checked regularly and shows no issues. I have also tried manually accessing the data and creating additional folders or some test data and everything works

Hi, see the solution here:

https://stackoverflow.com/questions/29244788/error-disk-i-o-error-on-a-newly-created-database

Does it not support cluster locking?

[Barlog9511 17]

3 hours ago, Luke said:

Hi, see the solution here:

https://stackoverflow.com/questions/29244788/error-disk-i-o-error-on-a-newly-created-database

Does it not support cluster locking?

NFS location in this case probably does not play a role, because this way it works for years and never had any problem and another thing is that I try targetnut backup of the rejoicing outside NFS on the root file system and the same error appeared
Some folders and subfolders were created in this location but the task still fired anyway

[Neminem 1518]

Have you looked at what it's backing up.

Please browse to the folders, and have a look.

Its 10*1000 files depending on your library size.

Each file text or images, 10kb to 10mb, Network I/O is a real problem when copying small size files over network.

This will help you a lot. 

I do zip them before network copy, 1 large file is faster than 10*1000 small files.

On 11/21/2024 at 7:13 PM, Neminem said:

I do local backups, and offload them to my backup nas, for more than 30 days of retention.

On my backup nas I have 6 month of rolling backups.

  

 

[Luke 42077]

12 hours ago, Barlog9511 said:

NFS location in this case probably does not play a role, because this way it works for years and never had any problem and another thing is that I try targetnut backup of the rejoicing outside NFS on the root file system and the same error appeared
Some folders and subfolders were created in this location but the task still fired anyway

I think it is related because SQLite has difficulty with nfs, and that’s the change made in the last year is that we’re using SQLite backup functions where we weren’t before. Did you at least like into what I linked to?

[Barlog9511 17]

27 minutes ago, Luke said:

I think it is related because SQLite has difficulty with nfs, and that’s the change made in the last year is that we’re using SQLite backup functions where we weren’t before. Did you at least like into what I linked to?

It does not help at all

Like I said in a previous message, I already tried backup on the root file system and the same error appeared

[Luke 42077]

What is the root file system?

[Barlog9511 17]

Debian server 12 with SSD as main HDD

 

[IAmHugh 63]

I attempted to setup backups yesterday and I'm having the same issue. I believe in my case that it may be due to the fact that in Arch /var/lib/emby is a locked directory unless I stop the Emby service or I access it as admin.

embyserver.txt

Edited November 28, 2024 by CummingCowGirl

[Luke 42077]

56 minutes ago, CummingCowGirl said:

I attempted to setup backups yesterday and I.m having the same issue. I believe in my case that it may be due to the fact that in Arch /var/lib/emby is a locked directory unless I stop the Emby service or I access it as admin.

embyserver.txt 49.84 kB · 0 downloads

Yes you are right:

	Application path: /usr/lib/emby-server
	System.IO.IOException: System.IO.IOException: Read-only file system : '/mnt/My Files I/Docs/Backups/Emby Backups/embyserver-backup-full'

 

[IAmHugh 63]

Is there any way around this? Maybe by adding permission(s) without changing the current permissions?

[Q-Droid 989]

In Arch you have to extend the service unit to allow read/write access to other filesystems/paths.

For example:
[Service]
ReadWritePaths="/mnt/My Files I/Docs/Backups/Emby Backups"
 

https://wiki.archlinux.org/title/Emby

[IAmHugh 63]

Looks like it it just might be easier to stop the emby service, manually backup, and then restart the service.

[IAmHugh 63]

I just tried systemctl start emby-server.service and checked /var/lib/emby/ and still locked. Shut down the server from within Emby and the folder is still locked. Any ideas?

[Q-Droid 989]

What do you mean by locked? What is it that's actually locked?

In the logfile you posted the errors are due to the backup destination being read-only for the Emby runtime user. The server can't create backups in that path. There's no indication of a problem with /var/lib/emby.

 

[IAmHugh 63]

All three of my external drives are setup the exact same way. GPT device table with  BTRFS partitions, so the permissions would be the same across the board. Why should something like backup for Emby have a issue accessing any folder on any of those drives when Emby accesses my media folders without issue? As for the locked folder.

First image Dolphin opened in my user account. Second image opened in root.

 

 

Edited November 30, 2024 by CummingCowGirl