Hi.
We have a Proxmox VE with ~50 CT's running.
We need to share some files between all the CT's. For that we use a NFS mount that must be mounted by all the CT's.
At the beginning I opted for mounting the NFS mounts in the host and then give them as resources using the -mpX parameter. I assume that this was a better approach, for the host kernel is just one NFS mount. I was happy with that solution.
But now I have added a second VE to the cluster.
I found that when I try to migrate a CT between servers in the cluster fails because the -mp resource.
If I remove the mounts from the container configuration and mount then inside the container with a mount call, the I can migrate the CT's without problem.
I just wonder if there's a way to make the NFS mount shareable between the two VE's.
Also wondering if mounting 50 NFS in 50 containers can have a negative impact in the host VE performance.
I will be grateful for any advise on this.
Cheers
Javi
We have a Proxmox VE with ~50 CT's running.
We need to share some files between all the CT's. For that we use a NFS mount that must be mounted by all the CT's.
At the beginning I opted for mounting the NFS mounts in the host and then give them as resources using the -mpX parameter. I assume that this was a better approach, for the host kernel is just one NFS mount. I was happy with that solution.
But now I have added a second VE to the cluster.
I found that when I try to migrate a CT between servers in the cluster fails because the -mp resource.
ERROR: migration aborted (duration 00:00:00): cannot migrate local bind mount point 'mp3'
If I remove the mounts from the container configuration and mount then inside the container with a mount call, the I can migrate the CT's without problem.
I just wonder if there's a way to make the NFS mount shareable between the two VE's.
Also wondering if mounting 50 NFS in 50 containers can have a negative impact in the host VE performance.
I will be grateful for any advise on this.
Cheers
Javi