Rebecca Mahany-Horton
Rebecca Mahany-Horton
This has been addressed with https://github.com/kolide/launcher/issues/896 and https://github.com/kolide/launcher/issues/954 -- the new autoupdate library lives outside of the bin path (i.e. in /var/..., outside of the nix store)
@lucasmrod thanks for looking into this -- I agree that this looks like it was fixed by locking client access. Probably good to close this out?
@directionless updated first bullet point accordingly
actionqueue should perform retries for failed actions without being prompted by control data changes
I wonder if just returning an error from `func (aq *ActionQueue) Update(data io.Reader) error` if any action fails to process would do it? Then the updated data for that subsystem...
NixOS support is newly available via https://github.com/kolide/nix-agent, with setup instructions [here](https://github.com/kolide/nix-agent?tab=readme-ov-file#running-kolide-launcher).
When testing I noticed the same behavior as James, where it takes about 10 seconds but does completely exit.
We added some additional logging [here](https://github.com/kolide/launcher/pull/1130) to try to understand what's causing the issue, and just got our first reproduction of the issue after releasing those changes. We saw this...
Saw this again after the release to v1.6.1 today
> Hrm... Under what circumstance does it take 5s, and if that happens, is it indicative of an issue? @directionless a lot of the time I see executable checks time...
I see this on a CentOS 7 VM also. systemd version is 219 -- looks like this was maybe fixed sometime after version 245 per https://github.com/systemd/systemd/issues/15275 Per https://repology.org/project/systemd/versions I tried...