3 ms·

That this was written by a (ex?) Red Hat employee makes a lot of sense. We burned significant time in the last 2 weeks because RHEL 8 intentionally makes it ver

That this was written by a (ex?) Red Hat employee makes a lot of sense. We burned significant time in the last 2 weeks because RHEL 8 intentionally makes it very hard to `dnf install docker-ce` (even though you are literally just installing the centos/rhel 7 package), and insteadm RH docs + staff tell IT admins who don't know better to `dnf install podman` as the only correct option and lie that is drop-in compatible. Then we get support tickets because, surprise, podman often doesn't actually replace docker due to bugs, or in areas like `docker-compose`, because it isn't even implemented. Docker's second death is very much IBM/ RHEL's anti-competitive intent. An infrastructure provider should be neutral on how they present this kind of thing, especially a company selling enterprise reliability for running software on top. The 1984-esque double speak experience has been quite souring wrt ethics of doing business with them around cloud/container-era technology and trusting them for enterprise reliability. I do find the podman ecosystem technically interesting... but due to the unethical corporate stewardship we're seeing, I'm uncomfortable seeing anyone use it.

tariqislam_ 6y ago
Author here. Ex Red Hat, yes. Couple of things: 1. Docker being deprecated has nothing to do with anyone's competitive intent. I believe that the Kubernetes community is optimizing away from bloat. 2. I do not espouse podman. This post was about the container runtime, which podman is most definitely not. It's been a few years since I left the company, but it's unfortunate that you've had the experience you describe.
- lmeyerov 6y ago
  I wouldn't have written anything if dockerless was opt-in for when you wanted to experiment with dropping `dnf install docker-ce`. There's a difference between infra orgs promoting new tech as an alternative vs. blocking their competitors. Infra co's are paid to be reliable neutral providers & advisors - instead, it seems corporate greed here has folks actively breaking enterprise userland and misleading decisions makers in sensitive areas like gov, utilities, world-scale services, etc. Like I said, I found the technical goal interesting and have been following. However, working through the year+ papertrail of RH doublespeak here makes me dread having to interact with them as an active OSS dev and in my dayjob working with big businesses / govs. RHEL often has policy lock-in in these orgs, so I know it'll keep happening unless the community stops it. As is, the stewards are needlessly risking putting podman and friends into a politically contentious category similar to urbit, oracle, etc., and are costing a lot of time & $$$ from people who are already stretched thin.
- mekster 6y ago
  I'm a random docker user but that is the same feeling. podman just doesn't work like docker and podman-compose isn't even at a usable state when RH deprecated Docker which seems completely rushed. And when you see that it's actually not so easy to install docker anymore, you get the feeling RH has another intention such as trying to control container business by themselves by trying to move people off docker. If it was purely from technical superiority, then it should've had a smoother migration path.

tariqislam_ 6y ago

Author here. Ex Red Hat, yes. Couple of things: 1. Docker being deprecated has nothing to do with anyone's competitive intent. I believe that the Kubernetes community is optimizing away from bloat. 2. I do not espouse podman. This post was about the container runtime, which podman is most definitely not. It's been a few years since I left the company, but it's unfortunate that you've had the experience you describe.

lmeyerov 6y ago
I wouldn't have written anything if dockerless was opt-in for when you wanted to experiment with dropping `dnf install docker-ce`. There's a difference between infra orgs promoting new tech as an alternative vs. blocking their competitors. Infra co's are paid to be reliable neutral providers & advisors - instead, it seems corporate greed here has folks actively breaking enterprise userland and misleading decisions makers in sensitive areas like gov, utilities, world-scale services, etc. Like I said, I found the technical goal interesting and have been following. However, working through the year+ papertrail of RH doublespeak here makes me dread having to interact with them as an active OSS dev and in my dayjob working with big businesses / govs. RHEL often has policy lock-in in these orgs, so I know it'll keep happening unless the community stops it. As is, the stewards are needlessly risking putting podman and friends into a politically contentious category similar to urbit, oracle, etc., and are costing a lot of time & $$$ from people who are already stretched thin.
mekster 6y ago
I'm a random docker user but that is the same feeling. podman just doesn't work like docker and podman-compose isn't even at a usable state when RH deprecated Docker which seems completely rushed. And when you see that it's actually not so easy to install docker anymore, you get the feeling RH has another intention such as trying to control container business by themselves by trying to move people off docker. If it was purely from technical superiority, then it should've had a smoother migration path.