This is what I'm looking for a quite long time. Especially with "crm configure" section starting from18:16. Thank you and appreciate you for this video. This is really helpful to achieve carbonio zextras email server with DRBD HA Cluster
Hey great video. One note: in corosync configuration, since you are using udp unicast option (udpu) I think you can further simplify your configuration file and remove redundant options 'mcastport' and 'broadcast'.
When I use the primitive fs_res ocf: heartbeat:Filesystem params ....till end command. It gives me this error: ocf:heartbeat:Filesystem got no meta data does this RA exist? Help is much appreciated
Thank you, Great Video It is working for me perfectly Can you please share any blog link if we want to attach this HA NFS cluster using bindnetaddr to Web Servers Facing issue to use this bindnetaddr
hi i need help you when i add the resource DRBD, the exit is: Error: Unable to create resource 'ocf:linbit:drbd', it is not installed on this system (use --force to override)
thanks for the video theurbanpenguin, but with the file configuration corosync.conf i have the nodes offline,both of them,i think i need to change something. im in ubuntu 18.04
So what if you run Nextcloud on this shared storage and managed as this setup. And then you upload many files, during the upload, the master reboots. This switches the roles of the slave to master. Then the slave (which is now the new master) is the one that can write stuff. But the old master still had some files on disk, not yet synced through DRBD. Is there then a problem with figuring out which files to keep or not? Since the DRBD master is what it says, it's the leading decision maker of what there is to sync, if this role switches a lot while you have unsynced files... Are some of these files then thrown away? Of course in this example, the client can resync the files it has to the cluster. But if this was a mailserver, then you don't want to end up with disregarded mail. Does this question make sense?