Benjamin Gilbert
Benjamin Gilbert
@tonglil As of Container Linux 1535.0.0, locksmithd defaults to performing an uncoordinated reboot after an update, which should handle your use case. `locksmithctl reboot` still assumes etcd, which is the...
Hmm, this introduces a divergence between files written with `-o` and those written by redirecting stdout. When running Butane from a container, we've generally recommended the latter. Does the divergence...
Our docs have always worked that way. If you want AMI IDs for older versions, see e.g. https://stable.release.core-os.net/amd64-usr/2079.6.0/coreos_production_ami_all.json.
We're not planning to update Container Linux to etcd 3.4.x.
Thanks for the report. Is there a previous Container Linux version where you weren't seeing this problem?
We generally support only low-level operations in the Ignition schema directly, and implement higher-level configuration as FCC sugar. This is a great example of something we may want to have...
Hi @netboy88, could you provide the information requested above? Otherwise there's not enough here for us to act on.
@netboy88 Thanks for the additional information. What behavior did you expect, and what happened instead?
Did older versions of Container Linux boot at a reasonable speed? Is there a particular part of the boot sequence that takes most of the time?
@vielmetti This seems likely to be a kernel bug (with an outside chance of networkd) and we ship a [pretty vanilla kernel](https://github.com/coreos/linux/pull/268/commits), so I wouldn't expect this problem to be...