4 ms·
The lesson is: don't build a system that requires or advertizes high avilability but that runs in a single datacenter. Load balance. Elect leaders. Batch, qu
by ojiikun 14y ago
The lesson is: don't build a system that requires or advertizes high avilability but that runs in a single datacenter. Load balance. Elect leaders. Batch, queue, and distribute. Fail-over. Fail fast. So long as we have earthquakes, lightning, human malice, and human error, single AZs will fail.
- corford 14y agoLooks like multi AZ setups are also failing :( http://news.ycombinator.com/item?id=4684491 http://news.ycombinator.com/item?id=4684491
- pjscott 14y agoRemember the old saying about "Lies, damn lies, and availability zones"?
- MicahWedemeyer 14y agoWe're paying 2x for the RDS multi-AZ setup that is supposed to failover in these cases. So far no luck. Think I'll ask for a refund.
- bryanh 14y agoSame here! Even trying to boot up some replicas is painfully slow.
- dangrossman 14y agoI gave up on Amazon (and ate the reserved instance fees) 2+ years ago over this. I was paying through the nose for multi-AZ RDS and when EBS failed in one zone, nothing failed over.
- evanmoran 14y agoWhat did you switch to?