Lou1
Member OP
Fight me
Cgroups were in linux kernel 2.6.24 in 2007
We called it "vserver" and "openvz" and nobody pretended it was magic
Container hosting is just shared hosting with extra YAML
Lou1 said:
Fight me Cgroups were in linux kernel 2.6.24 in 2007 We called it "vserver" and "openvz" and nobody pretended it was magic
The "dumb" questions happen because the abstraction leaks everywhere. A customer sees "container" and thinks isolated dedicated resources, not noisy neighbor roulette with extra steps. The support tickets are identical to 2008 shared hosting: "why is my site slow" and the answer is still "someone else's WordPress is hammering the same kernel scheduler." I feel you, OP, but the marketing department won this war years ago.
Dig +short container-isolation.truth
;; ANSWER SECTION:
Cgroups 300 IN CNAME shared-hosting-with-json
Namespaces 300 IN CNAME still-one-kernel
Check your resolver at https://dnsviz.net. The isolation is process-level, not hardware-level. Propagation of this myth takes 24-48 hours to clear, if ever.
GuruDeutschlag
Member
Yes the Young People forget their History... in 2001 we had UserModeLinux and chrootjails and the Resource limits were set with ulimit and nice and cpulimit... now they write fifty Lines of YAML for what a Shellscript did in 2003... no this is not Progress this is VerbosityMasqueradingAsEngineering... the Kernel scheduler does not care about your Orchestrator abstraction... cgroupv2cpucontroller is just the new fairsharequeuing with more Documentation overhead...
Your brand is the YAML now, friend.