Async

[johanot]

No description provided.

[srhb]

I pushed some commits. I think it eventually works most of the time now ™️

• I think the touch-rewrite was only half-done, think i fixed it in a sensible way.
• I fixed metrics so it doesn't spawn a runtime (inside the runtime)
• I enabled the secondary vault monitor config test.

There's some issues I am still debugging before I think it's ready for review by someone wiser than me. Writing them down here so I can keep track of where I am.

1. something is spamming

vm-test-run-faythe-vault-test> client # [   32.872869] faythe[881]: [2024-10-18T08:52:39Z ERROR rustify::client] error=Server returned error
vm-test-run-faythe-vault-test> client # [   32.873456] faythe[881]: [2024-10-18T08:52:39Z ERROR rustify::endpoint] error=Server returned error
vm-test-run-faythe-vault-test> client # [   32.874041] faythe[881]: [2024-10-18T08:52:39Z ERROR vaultrs::api] Detected errors in API response: []
vm-test-run-faythe-vault-test> client # [   32.874639] faythe[881]: [2024-10-18T08:52:39Z ERROR vaultrs::kv2] error=The Vault server returned an error (status code 404)

Haven't found out what yet. Probably part of the raw cert check.

2. faythe appears to set up too many challenges according to pebble. I've seen 5 challenges in the db at one time.

3. faythe appears to set up the wrong dns proofs sometime, getting stuck validating internally. Unsure why this happens.

4. sometimes there's nonce errors in the AcmeClient. I think this may simply be because of the default pebble setup vm-test-run-faythe-vault-test> acme # [ 5.144768] pebble[662]: Pebble 2024/10/18 08:52:12 Configured to reject 5% of good nonces -- but we should handle this (and maybe that'll fix some other issues? I do think I'm seeing them too often for the 5% to be the only reason)

[srhb]

I've fixed the remaining known bugs in our fork of acme_lib (tmp commit added while the fork change gets approved) -- at least I can now consistently run the test without failures...