Jump to content

tar -C fails


Daxten

Recommended Posts

tar -xf /mnt/user/appdata/gitlab/backups/repositories/xxx/xxx.bundle -C /mnt/user/appdata/gitlab/repositories/xxx/xxx.git

tar: ./hooks: Cannot utime: No such file or directory

tar: Exiting with failure status due to previous errors

 

works if I user disk1 or disk2 as a target, can't restore gitlab backup because of this. Using latest stable release

Anyone had this issue and knows workarounds / when this will be fixed?

 

Link to comment

which links? Can you point me to some docs or post?

is this unraid related or docker? :(

 

The symlinking that occurs when Unraid aggregates the shares.  So /mnt/user/share = /mnt/disk1/share + /mnt/disk2/share + /mnt/disk3/share etc

 

It's Unraid related but also affects some docker containers if you try and use /mnt/user/appdata/app rather than /mnt/cache/appdata/app

Link to comment

so, there is no way for me to scale gitlab over more then 1 disk?

 

This sounds pretty major to me, sucks if the devs really won't work on this :/

 

Well I'm not sure anyone has tried scaling gitlab over multiple disks, and I'd always assumed it was a limitation of Linux and the app you use not supporting symlinks but I may be wrong.  Got to ask though, how big is your gitlab?!  :o

Link to comment

so, there is no way for me to scale gitlab over more then 1 disk?

 

This sounds pretty major to me, sucks if the devs really won't work on this :/

 

Just looked at the command you're running, if it's all in your appdata share, then that should be cache only, so why can't you use /mnt/cache/ instead of /mnt/user/ ?

Link to comment

Just looked at the command you're running, if it's all in your appdata share, then that should be cache only, so why can't you use /mnt/cache/ instead of /mnt/user/ ?

My appdata is not cacheonly, which is not best-practice as it seems. I also have jira running on it and don't want to have all the files uploaded there on the cache (which may be worked around by adding an extra -v for the data in jira)

Link to comment

Just looked at the command you're running, if it's all in your appdata share, then that should be cache only, so why can't you use /mnt/cache/ instead of /mnt/user/ ?

My appdata is not cacheonly, which is not best-practice as it seems. I also have jira running on it and don't want to have all the files uploaded there on the cache (which may be worked around by adding an extra -v for the data in jira)

 

Yeah, it's not a great idea for the reasons I've specified above.  It will break some containers.

Link to comment

as a software dev myself (with not a lot of filesystem knowledge) I just ask myself why they wrote this awesome filesystem if it doesn't work like a filesystem? :P kinda destroys the purpose in my opinion, but this might be a special case and only my opinion

Link to comment

as a software dev myself (with not a lot of filesystem knowledge) I just ask myself why they wrote this awesome filesystem if it doesn't work like a filesystem? :P kinda destroys the purpose in my opinion, but this might be a special case and only my opinion

 

Yeah, I'm just struggling to see the advantage of having the data spread over disks, so don't really see why it's a problem.  Also like I said earlier I'd always assumed it was a Linux problem rather than an Unraid specific issue myself.

Link to comment

as a software dev myself (with not a lot of filesystem knowledge) I just ask myself why they wrote this awesome filesystem if it doesn't work like a filesystem? :P kinda destroys the purpose in my opinion, but this might be a special case and only my opinion

 

It's my understanding that the User Share system (/mnt/user and /mnt/user0) is built on FUSE, uses it to build virtual file systems over the top of the disk file systems.  You may want to research how FUSE works.  It's probably why links can't work.

Link to comment

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...