Twitter | Search | |
Derek Carr
232
Tweets
143
Following
322
Followers
Tweets
Derek Carr Dec 14
Replying to @derkoe @openshift
This is what I needed! Looks like it uses cpu sets. We are working to graduate support for that in Kube 1.10 but it will only help nodes configured with static cpu policy.
Reply Retweet Like
Derek Carr Dec 14
Replying to @derkoe @openshift
Unable to find crisp description of what is used by jvm when looking at “cpu limits”. Is it cpu sets or cfs quota? Many kube users do not set a limit and just a request which means cfs shares is all that is applied in cgroupfs. Details matter here.
Reply Retweet Like
Derek Carr Dec 13
If you run java apps on , I am interested to learn how you are configuring the jvm relative to pod resource requirements to know what we could do better.
Reply Retweet Like
Derek Carr Dec 11
Replying to @runc0m
Isn’t minikube integration done?
Reply Retweet Like
Derek Carr retweeted
Red Hat OpenShift Dec 7
Controlled chaos: Maturing Kubernetes weighs choice and standards (w/ on )
Reply Retweet Like
Derek Carr retweeted
ɹǝʞʎds ʍǝɹpuɐ Dec 6
Brave demoing CRI-O at live. Important focus on runtimes that are purposely built for just what orchestration platform needs.
Reply Retweet Like
Derek Carr Dec 6
Can make a point of securing me one of those?
Reply Retweet Like
Derek Carr Dec 4
Replying to @vishnukanan
The talk has a great title!
Reply Retweet Like
Derek Carr retweeted
Vishnu Kannan Dec 4
I'll be driving a dev summit session on scaling k8s for new contributors tomorrow. If you have a suggestion or feedback on that topic please reply. RT for visibility!
Reply Retweet Like
Derek Carr Nov 26
It’s hard to tell, but it looks like the cpu integration only works with cpuset and not cfs quota. Is there a good write up that says what exactly is read in cgroupfs? FYI static cpu pinning is alpha in Kube 1.8+ via cpuset
Reply Retweet Like
Derek Carr Nov 26
Any chance you could make the slides public?
Reply Retweet Like
Derek Carr Nov 20
Replying to @sarahnovotny
Pretty sure first mover can guarantee a win!
Reply Retweet Like
Derek Carr Nov 16
Replying to @gabrtv @jessfraz
we have work to do on the node to support in-place pod resource requirement updates, but yes, best effort is least effort
Reply Retweet Like
Derek Carr Nov 15
Looks cool!
Reply Retweet Like
Derek Carr Oct 26
Haha
Reply Retweet Like
Derek Carr Oct 22
When a cronjob created 15k jobs, I was motivated to quota all the things. Hope to finish it this week.
Reply Retweet Like
Derek Carr Oct 19
I love that I got some Kubernetes socks from this talk.
Reply Retweet Like
Derek Carr Oct 19
I just saw 9th floor was renovated yesterday. Thought I was on wrong floor!
Reply Retweet Like
Derek Carr Oct 5
I volunteer eparis
Reply Retweet Like
Derek Carr Oct 5
changing a log level on the fly would be AMAZING
Reply Retweet Like