[DRBD-user] Control deployment?

Frank Rust f.rust at tu-braunschweig.de
Thu Nov 2 16:01:24 CET 2017

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.

> Am 02.11.2017 um 15:00 schrieb Yannis Milios <yannis.milios at gmail.com>:
>>> (Now I am solving this by migrate vm to node1, unassigned vm-image from node3, assign vm-image to node 3, migrate vm to node3, unassigned vm-image from node1, whats awful, error prone and somewhat waste of time)
> Why you are doing all these steps manually?! This is something that PVE handles automatically via the DRBD9 plugin. Just a simple offline/online live migration of VM from one node to the other should be enough to set drbd resources in the correct mode…

Yes, I thought so. But it does not…  
When migrating the vm from node3 to node1, the diskless assignment on node3 stays (secondary), the one on node1 becomes primary. And when moving the vm back to node3 it keeps the diskless and just makes it primary again. 
And when I migrate the vm away, remove the diskless assignment, and move the vm back, then it is recreated diskless…
Of course this makes sense, not to copy the vm-images around, when moving the vm. 

Possibly there should be a (proxmox) module to do "intelligent" deployment. By now it just takes the largest free space(s) to place the volumes.
Until then I want to do smart things by myself, but I am failing too. 

Is there any documentation to create deployment plugins?

Best regards,

More information about the drbd-user mailing list