We're having issues with AWS Aurora for MySQL. Aurora uses Route 53 for DNS load balancing amongst reader nodes. When diagnosing by doing nslookup to the RDS cluster hostname it will sporadically return a public IPv4 address and not an address inside the VPC. We don't allow public ingress to RDS so the connection times out.
Anybody else confirm this is affecting RDS? I have an RDS PostgreSQL instance that has been stuck in the deleting state for over 60 minutes. I love (not) how the notice in the AWS control panel does not list RDS as an affected service.
The resolution issue we were seeing has subsided as of about 16:01:30 UTC-7 and resumed again at 16:08:00 UTC-7. I haven't tried deleting any instances, so can't confirm on that particular issue. Hopefully they get this cleared up quick- this is a nasty issue. I can't create new reader nodes and can't hardcode an IP address as the traffic requires multiple readers.
3 comments
[ 2.7 ms ] story [ 11.9 ms ] thread