Hi,
First of all, I know it is not recommended but I feel like my scenario should work?
EDIT:
Ok, ventured into the bowels of the beast and I am very close to the issue.
In https://github.com/proxmox/pve-container/blob/master/src/PVE/LXC.pm#L1218 all mount points are send somewhere?
My dirty little hack is now to simply strip sshfs from the list before it gets send:
It works but I still have no clue what the issue is, as I do not speak perl, I fail to trace where this mounts list is send to.
I mean, I see it is a process inside the container? but said container is frozen, so I can not poke there... (see my reply).
I have an lxc container on pve 7.1-8 with a separate mount point:
and disabled backups on it.
I have mounted an sshfs share inside the container into /ext/sshfs.
I use snapshots as backup mode onto a pbs.
Creating a backup stalls indefinitely trying to create a snapshot:
/ext is excluded:
and due to /ext being a separate mount point it is a separate zfs dataset:
so how can vzdump freeze if it is not even supposed to touch /ext?
(I have tried several no-cache options for sshfs:
but nothing worked).
So, besides it not being recommended, shouldn't this work?
I feel like this is a bug in vzdump as it seems it is not leaving my mountpoint alone?
Thanks for any feedback!
First of all, I know it is not recommended but I feel like my scenario should work?
EDIT:
Ok, ventured into the bowels of the beast and I am very close to the issue.
In https://github.com/proxmox/pve-container/blob/master/src/PVE/LXC.pm#L1218 all mount points are send somewhere?
My dirty little hack is now to simply strip sshfs from the list before it gets send:
Perl:
$mountdata =~ s/.*fuse.sshfs.*\n//g;
I mean, I see it is a process inside the container? but said container is frozen, so I can not poke there...
I have an lxc container on pve 7.1-8 with a separate mount point:
Code:
local-zfs:subvol-107-disk-1,mp=/ext,size=1G
I have mounted an sshfs share inside the container into /ext/sshfs.
I use snapshots as backup mode onto a pbs.
Creating a backup stalls indefinitely trying to create a snapshot:
Code:
INFO: create storage snapshot 'vzdump'
/ext is excluded:
Code:
INFO: including mount point rootfs ('/') in backup
INFO: excluding volume mount point mp0 ('/ext') from backup (disabled)
and due to /ext being a separate mount point it is a separate zfs dataset:
Code:
rpool/data/subvol-107-disk-0 3.38G 1.62G 3.38G /rpool/data/subvol-107-disk-0
rpool/data/subvol-107-disk-1 96K 1024M 96K /rpool/data/subvol-107-disk-1
(I have tried several no-cache options for sshfs:
Code:
cache=no,noauto_cache,entry_timeout=0
So, besides it not being recommended, shouldn't this work?
I feel like this is a bug in vzdump as it seems it is not leaving my mountpoint alone?
Thanks for any feedback!
Last edited: