Add retries to iPXE script when fetching files

[mmlb]

Add retries to iPXE script when fetching files

[tstromberg]

Can someone identify which line in boots is responsible for the fetching that is missing retry?

It would help the person who might want to take this issue on.

[jacobweinstock]

If I'm not mistaken, @mmlb is talking about this area in job/ipxe.go, which is in the function serveBootScript.

[mmlb]

Nope, its actually in the installers: https://github.com/tinkerbell/boots/blob/ba3a3fef424ebfd7125b08ae99dcb9631bc911a8/installers/osie/main.go#L54 We'd probably need a bounded loop that breaks on success.
And similar for fetching the initrd and posting events back.

[tstromberg]

If it helps, I've had good luck with using this library for exponential backoff+jitter, and have seen it used elsewhere in Tinkerbell (tink?): https://pkg.go.dev/github.com/cenkalti/backoff/v4

[mmlb]

The http request is not from boot's side. Its ipxe doing the fetches we'd need to do the retries/backoff in the iPXE script.