I'm seeking advice on an issue with Kubernetes, specifically regarding how to reliably capture and store large memory crash dumps (over 100GB) from a Windows pod running in Azure Kubernetes Service (AKS) after a crash. It's crucial that these dumps are saved without any corruption so that I can download or inspect them later. I've experimented with a premium Azure disk (az-disk), but it hasn't been reliable for this situation. I'm also exploring options like emptyDir, though I haven't tried that yet. Any suggestions would be greatly appreciated!
2 Answers
Are you using these dumps mainly for debugging? Understanding the use case can really help in advising the best approach. And yeah, dealing with Windows containers is not the most fun—I've been there!
Yep, definitely debugging! I just wish these Windows pods weren't so tricky in Kubernetes.
I feel you on the large dumps! Just a heads up, make sure to check your overall application design. If you're consistently running into large memory dumps, maybe there's a way to refactor your app to manage memory better. I learned the hard way with mine!
Yeah, plus debugging those huge dumps can be such a hassle!
Right? That was my first thought too! Let's hope they get the help they need!
Same! They can be such a pain compared to using Linux containers.