Travis Nielsen

@drohnow Rook has a crash collector daemon that watches for the dumps and reports them to ceph, so they can be viewed from the toolbox. Do [these commands](https://docs.ceph.com/en/latest/mgr/crash/) in the...

Do you see a crash collector pod already running? It should have collected the crashes for you and reported them to ceph. From the toolbox, do other commands such as...

Ok great, so that answers your question? You can also get details about each crash with `ceph crash info `.

hmm, the backtrace is certainly not helpful. All of the traces are this unhelpful?

@gauravsitlani Could you share your insights to gathering the core dump?

I wonder if we should always enable the needed property so the core dumps would be enabled by default. But at the same time, `ShareProcessNamespace` and `HostPID` are significant privileges...

CIS - 5.4.1 states that it is preferred to use files instead of env vars in case the pod were to inadvertently log its environment. While this is a valid...

Could you remove the resource limits on the OSDs (or increase the memory/cpu to some really high value)? Since you're seeing the change in perf over time with the same...

To capture scenario from the slack discussion... the scenario is where each datacenter would have its own storage class, instead of following the pattern of cloud providers that have a...

Still an interesting scenario