Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On 11/18/11 16:56, Miles Fidelman wrote: > Hi Folks, > > I've been running a 2-node HA cluster for several years - Xen VMs backed > by DRBD volumes + pacemaker for failover. > > I've recently added 2 newer nodes, and I'm trying to figure out the most > straightforward way to migrate a couple of production VMs from the old > hardware to the new hardware - with minimum outage. > > Just to complicate things a bit, I'm running DRBD 8.0.14 - which doesn't > support a 3-node configuration. > > Any suggestions re. a sequence of operations that allows me to: > 1. install latest versions of everything (Debian, DRBD, Xen, Pacemaker, > etc.) on the new nodes (easy) > 2. replicate data volumes to the new nodes while the VMs continue to > run on the old nodes > 3. migrate the VMs to the new nodes > 4. decommission the old nodes > > - all with minimal downtime for the VMs, and minimal (or no) period > where I don't have a backup copy of the data? This is assuming you're running your existing Pacemaker cluster on Heartbeat. How about: 1. Shut down one of the old nodes, have Pacemaker take over all cluster services to the other. 2. Bring up the first of the new nodes. Do not start cluster services yet. 3. Sync DRBD. DRBD 8.3. will happily talk to 8.0.14. 4. Put Pacemaker in maintenance mode. 5. Start Heartbeat. Heartbeat 3.0.4 will happily talk to whatever Heartbeat version you're using. So will Pacemaker. 6. Remove the old node from the Pacemaker configuration ("crm node delete"). 7. Take Pacemaker out of maintenance mode and move resources over. 8. Shut down the other old node. 9. Bring up the second of the new nodes. 10. Repeat steps 3-7. 11. (optional) Put Pacemaker in maintenance mode again, stop Heartbeat, replace with Corosync, disable maintenance mode. Does this help? Cheers, Florian -- Need help with High Availability? http://www.hastexo.com/now