Container replication stucked at "freeze guest filesystem"

jaysee2607

New Member
Aug 1, 2023
4
0
1
Hello,

During my test to HA, i'm trying to enable replication between 2 proxmox 8.0.3 nodes (using ZFS).

When lauching replication, i'm stucked at "freeze guest filesystem".

This only happens if option FUSE=1 is checked (and my CT has a SSHFS mounted point)

Unchecking the option (and rebooting the CT) and it works.

[EDIT] CT configuration:
arch: amd64
cores: 2
features: fuse=1
hostname: xxx
memory: 4096
net1: name=net1,bridge=vmbr3,gw=10.0.3.196,hwaddr=72:8A:31:09:E2:0A,ip=10.0.0.201/16,type=veth
onboot: 1
ostype: debian
rootfs: zfs:subvol-201-disk-0,size=500G
startup: order=90
swap: 512

Any clue?

thanks

[EDIT2] I got others CT with option FUSE=1 that work well... Actually I've got a bunch a servers with the stuck problem, all have more a 100Go of data, seem the zfs snapshot take a long time... normal ? what can I do to increase speed ? (doesn't explain why unchecking FUSE=1 fixed the thing)

[EDIT3] This only appears when the sshfs is mounted, if not mounted, works. my fstab for the sshfs point is :
xxx@xxx:yyy /yyy fuse.sshfs delay_connect,_netdev,transform_symlinks,identityfile=/ttt/id_rsa,allow_other,default_permissions,uid=0,gid=0 0 0
I tried some some more options, but without success (and for others CTs theses options are working!)
 
Last edited:
Hi Nico,

sorry as I didn't find any solution, I managed to find another solution avoiding usage of FUSE=1 and SSHFS. Then no problem.

Did you have any change disabling your FUSE (and any FUSEFS connection)??
 
This is a known problem with fuse in containers. Please use the fuse feature with caution.