Friday read --- scariest Docker in production story ever told

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

TuxDude

Well-Known Member
Sep 17, 2011
616
338
63
Yup - full of FUD. Just some dude ranting because he had some bad experiences, possibly because he doesn't know what hes doing. I was fining so many technical errors while reading it that I just had to stop and close that tab before I got to the end.
 

dswartz

Active Member
Jul 14, 2011
610
79
28
Not arguing with you, tuxdude. I can spell docker, but that's about all. What specifically was FUD and so on?
 

TuxDude

Well-Known Member
Sep 17, 2011
616
338
63
I'm not going to go back through it all with a fine-tooth comb, but just taking a quick scroll back through and looking mostly at the headings or otherwise bolded text (why is there so much bold anyways, me-thinks someone likes a bit of sensationalism)...

"Linux 4.x: The kernel officially dropped docker support"

Really - hes going to make that claim? Besides the fact that the kernel has never had "docker" support - the kernel provides several cgroup namespaces, docker requires most of them (as do LXC and most other container implementations), and none of them were dropped in kernel 4.x

But wait...., he didn't actually mean that they dropped docker support, below that it says:

"the AUFS filesystem was finally dropped in kernel version 4"

Which is also not true - AUFS was never merged into the upstream kernel. Can't drop what you don't have. It might have been dropped from the debian/ubuntu distributions around the time that they moved to 4.x kernels, I can't comment on that as I tend to stay away from those and prefer RHEL and the various clones. RHEL (and CentOS etc.) of course has NEVER included AUFS, yet has supported Docker for quite some time, which brings me to yet another unexplainably bold line:

"How does docker work without AUFS then? Well, it doesn’t."

Ya - tell that to the engineers at RedHat. The author of that article though seems to ignore the fact that distributions outside of the Debian-based group exist, with statements like:

"This issue is worldwide. It affects ALL systems on the planet configured with the docker repository."

Well.., the docker apt repo maybe. I won't argue whether they had some certificate issues with their apt repo, as I've never even tried to use it. But I can guarantee that any problems that it did have did NOT affect any RPM-based systems on the planet.



Overall - the article is so littered with things I know to be wrong, that I have serious doubts about every other point he makes, and so trust none of it. The whole thing just gets categorized as FUD.

As to my own experience with Docker, which I will admit does not involve running things at close to the scale that the author of that article does - docker can be great, when used correctly. You can NOT just migrate an app over to docker containers though - it is a whole different architecture and mindset around how to build/deploy applications. For almost all of the stuff I deal with at work (traditional enterprise apps like erp's and such) switching over to docker just doesn't fit or make any sense.
 

TuxDude

Well-Known Member
Sep 17, 2011
616
338
63
The retort is reasonable. I had low expectations after taking issue with the very first sentence, but then it went on to make some good points. But calling the original article "well written"...., thats a label I don't apply to rants.

It also has a great line I'm going to quote here, as I think it really applies to most of us well in our home-lab activities.

new and shiny does not bring home the bacon, plain and boring does.
We all love playing with the shiny new stuff - its what gets most of the talk around here. But as can also be seen around here, its the boring old stuff that we all rely on to "just work" that makes everything go, like good old pfSense and/or ZFS that underpin so many of our labs running the shiny and new.
 
  • Like
Reactions: dswartz