Hello,
We have received a new requirement to setup Connections' latest version with two DB nodes ( 1 app server + 2 HADR DB2 servers ) on DC and DR setup will have a single app node + 1 single DR DB2 node.
The database replication would be happening through the HADR method between two DC and DR environments ( 1 primary standby + 1 Principal standby at DC and 1 auxiliary standby DB at DR )
My query is that can we use the SCP or rsync commands to replicate the Content stores data from DC to DR and vice versa? or are there any other options of replication for the Content store with DR?
Will the data be in an inconsistent state if the replication through SCP / RSYNC tools in Linux?
Kindly share your inputs. Thank you in advance.
Regards,
Elango
Hi.
this is an interesting setup.
Have fun automating the process to get the DR environment live in case of the failure of the productive instance.
After a side-by-side migration, there are always so many steps to take until the new instance comes up without errors.
For the content store, we use Rsync to migrate the data to the new instance. Rsync has the advantage that it can compress during transfer, does incremental updates and deletions in the destination.
Maybe your NFS Server has some features to copy over the data. Unfortunately, I would not copy over every file in the NFS. I would spare out all static data and WebSphere cell-specific data like the MessageStorre files. Therefore a full NFS copy might not that good.
Do you already use the K8s infrastructure? What is your plan to synchronize this? I have no idea if Rsync works for persistent storage as it contains more and larger binary data than the standard NFS for the WAS infrastructure.
Martin
@Martin Thanks for your response.
We are planning to build this setup in the new Linux environment as there would not be any migration activity for DR. We will install and configure the Connections on both DC and DR environments.
This is not a k8s infrastructure, we will be doing this on CentOS/RHEL OS.
Do you have any idea regarding DR setup for Connections and data replication with DR?
I see two issues with syncing shared directory. Like Martin said there are the messageStores involved and I'm pretty sure they don't like this.
But I'm curious how do you switch the shared directory from one to the other for a failover case? Just unmount / mount on WebSphere? Pretty sure that this won't work without a restart, maybe more tweaking needed.
DB HADR is not an issue, but the file stores and messagestores will.
Cheers
Christoph