Hi!
In my constant chase to get out of my PLP NVMes and 10gb network I came to realise that ceph wasn't cutting it when it came to performance while reliability and having a proxmox GUI absolutely rocks. (Thats not even mentioning being able to SEE the files in Gluster even with a broken MDS/MON or OSD).
In order for me to replicate what I have on ceph to ZFS+Gluster I have done a POC where I got HA working just fine when both machines were up, getting much less speeds but still much better than ceph sadly.
That said, moving a VM to a machine I then shutdown makes it impossible to move from /etc/pve/qemu-server of the shutdown machine to the live one like on ceph.I shall try to put it in a HA group to see if I have better luck.
Any tips are welcome!
In my constant chase to get out of my PLP NVMes and 10gb network I came to realise that ceph wasn't cutting it when it came to performance while reliability and having a proxmox GUI absolutely rocks. (Thats not even mentioning being able to SEE the files in Gluster even with a broken MDS/MON or OSD).
In order for me to replicate what I have on ceph to ZFS+Gluster I have done a POC where I got HA working just fine when both machines were up, getting much less speeds but still much better than ceph sadly.
That said, moving a VM to a machine I then shutdown makes it impossible to move from /etc/pve/qemu-server of the shutdown machine to the live one like on ceph.I shall try to put it in a HA group to see if I have better luck.
Any tips are welcome!
Last edited: