Did anyone already implement Archer v6 (experiences with earlier versions are also interesting) with the AlwaysOn clustering feature of MS SQL (available since version 2012)? SQL clustering is mentioned in the sizing and performance guide but only at a very high level and without explicit mention to AlwaysOn.
I would be most happy to hear feedbacks, dos and don'ts, connectivity issues during fallback, etc.
Thanks in advance,
- Community Thread
- database cluster
- Forum Thread
- high availability
- ms sql
- RSA Archer
- RSA Archer Suite
- Version 6
I spun a v6.0 Beta instance on premesis here for our evaluation very early on. We ran bare bones system, so no AlwaysOn SQL, but I did go through the entire environment construction and set up to a working instance. However, given that it was a beta there were a lot of things I was unable to do/setup/configure/use, which have since been addressed.
Although, I have not had a chance to work on such a setup. I assume the failover in an AlwaysOn environment is quite transparent to the applications that are connecting to it. At the SQL level in a fail-over cluster the databases are replicated to the other nodes through Log shipping and when it does promote a secondary node to primary, Archer should be able to connect to it.
That being said, I think it would become more complex if you are setting up a backup instance or a disaster recovery environment because you will need to take care of:
- File repository
- Company Files
In a pure DR setup : The SQL Always on will take care of the database replication (Instance db) and each environment will have its own config database. Technically, you will have a DR environment waiting to start (i.e. connected to the replica of the primary db & archer services stopped). Once your primary Archer environment goes down, you will need to fire up the services in the DR to resume normal operation. But such a setup brings up the problem of the Search Index files - I'm not sure if copying the index files is good enough..
Thank you Syed and Scott. I was not considering a DR setup, here. I was more considering a pure HA environment (ie considering the DR is otherwise taken care of). I would be super interested in test results on cluster fallback. In theory Archer should reconnect "smoothly" but what "smoothly" means in practice, I'd like to see it with my eyes.
Technically the transition should be smooth. Worst case, I can see people being logged out if tblSession is not synched in realtime when the transition occurs. And I would like to find out if the indexes will be broken too.. That would be an interesting test.