Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
All, I am working on a strategy to upgrade my real time highly available application that uses a drbd resource as its file system for its database. I am using drbd 8.2.5 and heartbeat 2.1.3, on a CentOS 5.2 server pair. The idea goes something like this: 1. Set drbd on the Primary host to StandAlone, so that the real time app continues to run, but releasing the drbd resource on the Secondary host 2. Mount the drbd resource on the secondary, delete the database 3. Upgrade the real time app and its database on the Secondary host 4. Switch the running real time app to the Secondary host 5. Reconnect the drbd resource on the Primary host to the Secondary host, and resync from Secondary to Primary I've gone over the drbdadm tool, and I think the steps are something like: (on Primary) drbdadm disconnect pgsql (on Secondary) drbdadm primary pgsql; mount /dev/drbd0 /mirror Upgrade app and db on Secondary Switch running app to Secondary (on Secondary) drbdadm invalidate-remote pgsql (on Primary) drbdadm connect pgsql Can someone confirm that I've got the right commands and in the right order? Have I provided enough info? Its one of those things where I know I could probably hose the servers if I don't do it right, and I've pretty much got one shot to get it right. Thanks for your help, Doug Knight WSI Corp -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20080905/89c5dc6b/attachment.htm>