@@Mallik034 Many thanks for your reply. As update to my situation, it turned out that our Backup/Recovery (BUR) team's server filesystem backup jobs were cause for frequent node evictions across multiple environments (i.e., dev, test, prod) in which I support. Once BUR made adjustments to jobs, it seemed to have alleviated and evictions have ceased for past 4-5 weeks now. However, still have concerns with how our interconnects are setup, given they're all on the same subnet. Network engineers mentioned all interconnect IP using dedicated switch/routers, but I'm not convinced it's setup correctly per Oracle Support best practices.
Mallik...nice...but would be good if you add some more details.
Sure, I will review it once again and I will add more details explanation along with some technical demo. Please standby
@@Mallik034 Many thanks for your reply. As update to my situation, it turned out that our Backup/Recovery (BUR) team's server filesystem backup jobs were cause for frequent node evictions across multiple environments (i.e., dev, test, prod) in which I support. Once BUR made adjustments to jobs, it seemed to have alleviated and evictions have ceased for past 4-5 weeks now. However, still have concerns with how our interconnects are setup, given they're all on the same subnet. Network engineers mentioned all interconnect IP using dedicated switch/routers, but I'm not convinced it's setup correctly per Oracle Support best practices.
Thank you for sharing. Would you agree that private interconnect should be on it's own subnet and not be pingable from outside Oracle RAC cluster?
That's correct