[Podman] Re: Q re errors pulling a docker image

On 2020-04-28 13:36, Philip Rhoades wrote: > People, > > When I run this: > > # podman run \ > --publish 3085:3085 \ > -d \ > --name coda \ > codaprotocol/coda-demo:sandbox-32a.1 > > I get this: > > ERRO[0003] Error deleting network: missing prevResult from earlier > plugin > ERRO[0003] Error while removing pod from CNI network "podman": missing > prevResult from earlier plugin > ERRO[0003] Error refreshing container > 19e58fe9a7c81de33fb8833610cd9a0be3edadf867acf7acf9b3c78a32483835: > missing prevResult from earlier plugin > Trying to pull docker.io/codaprotocol/coda-demo:sandbox-32a.1... > . > . > > The container starts up OK and I am about to see if it is actually > doing what it is supposed to (I am just getting started with Coda > Protocol) but should I be concerned about those errors? This usually occurs only once after a system reboot, when Podman was not able to cleanly shut down containers. Most of the per-boot state is kept on tmpfs filesystems and will be wiped without intervention on reboot, but this is unfortunately not the case for IP address reservations in CNI, which for reasons I've never heard adequately explained live somewhere in /var and survive reboot. This error is the result of us (unsuccessfully) attempting to clean these reservations. In short, if it's only happening once per boot, the first time a Podman command is run, I would not be terribly worried. You're probably leaking one or two IP reservations per reboot, but the default network for containers is a /16, so it's very unlikely this will actually matter. Podman 1.9.0 and higher replaced the code that did this (which not very durable against failure) with a new method that is much more reliable, so this should not be a problem in the future.