Been a busy week and learned a lot.. However, the next thing I'm trying to tackle is being able to connect my storage and peripherals to my LXCs and VMs.
I was thinking I could maybe be creative and try to use the resource mapping feature for the USB Z-wave stick, but migration still complains that it can't migrate with it connected... Honestly, I could care less if this Z-wave stick survived the migration, as long as the VM made it over and started running again. I would still benefit from getting 95% of my system being functional.
Similarly, I was excited to try the directory mappings feature, but I realized that only applies to VMs and not LXC. I have set up each node with the proper NFS mount directory, which allows me to mount it to the LXC individually, but it prevents the container from migrating if it is mounted, which isn't ideal.
What's the best way to allow the use of shared storage and also either allow the usb stick to persist between nodes, or force the migration between cluster nodes, disregarding the mount point or usb hardware? It seems like migrations with either of those devices could work if proxmox did not error out (each node could just have the mounted directory for the NFS share, for example).
Currently both of my nodes are ZFS, but replicate via jobs. Migrations generally seem to work ok except if a VM or LXC have one of these illegal devices. For me to migrate, it usually forces me to remove the entry before it allows it to move. Are there any types of automated scripts that can run upon migration? My HA strategy is to migrate once without failback, as I would like to be able to check everything out before manually migrating things back.
I was thinking I could maybe be creative and try to use the resource mapping feature for the USB Z-wave stick, but migration still complains that it can't migrate with it connected... Honestly, I could care less if this Z-wave stick survived the migration, as long as the VM made it over and started running again. I would still benefit from getting 95% of my system being functional.
Similarly, I was excited to try the directory mappings feature, but I realized that only applies to VMs and not LXC. I have set up each node with the proper NFS mount directory, which allows me to mount it to the LXC individually, but it prevents the container from migrating if it is mounted, which isn't ideal.
What's the best way to allow the use of shared storage and also either allow the usb stick to persist between nodes, or force the migration between cluster nodes, disregarding the mount point or usb hardware? It seems like migrations with either of those devices could work if proxmox did not error out (each node could just have the mounted directory for the NFS share, for example).
Currently both of my nodes are ZFS, but replicate via jobs. Migrations generally seem to work ok except if a VM or LXC have one of these illegal devices. For me to migrate, it usually forces me to remove the entry before it allows it to move. Are there any types of automated scripts that can run upon migration? My HA strategy is to migrate once without failback, as I would like to be able to check everything out before manually migrating things back.
Last edited: