Hey experts, after your opinion here on these.
At my company we have a problem with windows clusters using regular always on stretched over multiple data-centres because the link between them is unstable (long distance). It often causes issues with production when the DR side experiences issues from both the windows cluster perspective and the log files growing in production.
We have tried to use distributed groups to resolve this, as the clusters can then be isolated per DC isolating the cluster failure. Which has worked to some extent.... however:
- I find the administration of distributed groups clunky, as no administration can be done through the GUI, seems like these were implemented as an afterthought my MS.
- Automatic seeding through distributed groups seems to be completely random to its success or failure.
- Monitoring these are difficult, I approached many monitoring providers and none seem to provide monitoring for this. Furthermore it was not even on their road-map.
- The only way to re-add the database on the secondary AG group seems to be to re-add it to the first one. There seems to be no process to repair the secondary from Microsoft. Automatic seeding doesn't self heal.
Does anyone use these in aggression? I was hoping for further improvements to these in SQL 2019 but from what I can see there doesn't appear to be.
I might need to ditch these and go back to regular always-on over a stretched cluster, which is a shame!
- This topic was modified 2 months, 2 weeks ago by dannyboy121.