This discussion has been locked. The information referenced herein may be inaccurate due to age, software updates, or external references.
You can no longer post new replies to this discussion. If you have a similar question you can start a new discussion in this forum.

Moving from Active-Active to Active-Passive

Looking for some thoughts please on the best way forward to move from a current active-active model to an active-passive model.

Currently we have an active-active model where we have seperate Solarwinds Servers / SQL Database Servers at a Primary and DR Site, with all of our network devices being polled by both sites.

ActiveActive.jpg

We want to move to an active-passive model using Solarwinds HA so that the Application servers at the DR site are in passive mode awaiting failover.

If we configure it like below we have a HA VIP with the data going to the current ACTIVE Solarwinds Server, followed by a SQL Cluster VIP again going to the current ACTIVE SQL Server.

However it appears that by doing it this way only one of the actual databases will be written to.

ActivePassive.jpg

This isn't exactly ideal in that the other database at the DR site isn't being written to / updated in any way.

If we were to lose the Primary site we would lose the database also, so we would need to find the best way to replicate the database data to the DR site at the same time to prevent this from happening if possible.

I know that Solarwinds currently don't provide HA support for the Database, but i need to find a way to keep the system functional with the minimum downtime possible in the event that the primary site ever goes down.

In our current active-active setup we have a process whereby we take the latest database backup from our primary site and restore this data to the DR database and make some changes to the database and then run the configuration wizard to point the DR instance to the DR database and we are back in action. The only reason for doing this is that most configuration changes to the database are only done on the primary site. Bad practice i know and double the amount of effort, hence why we want to move to the HA Active-Passive model.

Ideally i would like to be able to do the below but don;t know the best way to achieve this with the least amount of effort.

ActivePassiveReplication.jpg

Maybe it's wishful thinking on my part that the above is actually achievable, if it is then great, i need to figure out how to do it.

If however its not possible then i need to be able to somehow have a working system in the event of the primary site not being available.

Sorry for the long winded post, i appreciate any suggestions / reference docs that may be linked for my reading.

I'm surely not the only one thats wanting to have this kind of setup, others must be doing some sort of HA / DR config and would appreciate any help anyone can offer.

Thanks in advance.

  • SQL AlwaysOn Availability Groups should be used with HA to provide redundancy to the database and provide replication. This can be used in both a same-subnet, or multi-subnet failover configuration.

  • aLTeReGo​ Thanks for the reply, can you advise however which is better.

    Basic Availability Groups (SQL Standard)

    OR

    AlwaysOn Availability Groups (SQL Enterprise)

    Apart from the difference in licensing, are both methods supported?

    We only have a Solarwinds Database, we don't use NTA (yet).


    Regards

    Duncan

  • Basic Availability Groups (BAG) are sufficient for Orion's needs. If you have more than one database, such as in the case of NTA or Log Analyzer, you will need to create multiple availability groups since each BAG only supports a single database.

  • Can you please provide detail on how you are doing the following?

    "we take the latest database backup from our primary site and restore this data to the DR database and make some changes to the database and then run the configuration wizard to point the DR instance to the DR database and we are back in action."

    I am specifically interested in the "and make some changes" statement.  Currently I have two independant instances running.  And I would like to be able to perform this database restore action that you are stating above. 

    Thank you

  • Hey there,

    Sorry but just seen this message - give me a day or so and i will provide some info on what we did when we were in an active - active setup.

    Not sure though if your two installations are polling the same equipment like ours was, but it shows what we needed to do when we failed over to our DR box.

    Cheers

    Duncan