Closed travier closed 5 months ago
From the logs:
[2024-01-25T10:38:28Z INFO fcos_graph_builder] starting server (fcos-graph-builder 0.1.0)
[2024-01-27T07:30:00Z ERROR fcos_graph_builder::scraper] transient scraping failure: error sending request for url (https://builds.coreos.fedoraproject.org/prod/streams/testing/releases.json): error trying to connect: Connection reset by peer (os error 104)
[2024-01-27T07:30:00Z ERROR fcos_graph_builder::scraper] transient scraping failure: error sending request for url (https://builds.coreos.fedoraproject.org/updates/testing.json): error trying to connect: Connection reset by peer (os error 104)
[2024-01-27T13:58:29Z ERROR fcos_graph_builder::scraper] transient scraping failure: error sending request for url (https://builds.coreos.fedoraproject.org/updates/next.json): connection error: Connection reset by peer (os error 104)
[2024-01-28T06:17:46Z ERROR fcos_graph_builder::scraper] transient scraping failure: error sending request for url (https://builds.coreos.fedoraproject.org/updates/stable.json): error trying to connect: Connection reset by peer (os error 104)
[2024-01-30T06:39:28Z ERROR fcos_graph_builder::scraper] transient scraping failure: error sending request for url (https://builds.coreos.fedoraproject.org/updates/next.json): error trying to connect: Connection reset by peer (os error 104)
[2024-01-30T06:39:28Z ERROR fcos_graph_builder::scraper] transient scraping failure: error sending request for url (https://builds.coreos.fedoraproject.org/prod/streams/next/releases.json): error trying to connect: Connection reset by peer (os error 104)
Made https://github.com/coreos/fedora-coreos-cincinnati/pull/95 as an option to push an update.
If you look at the timestamps in the Cincinnati logs, they're few and far between and represent transient failures.
The Zincati logs appear more frequent but stopped after January 25th (which is also near when Cincinnati was last restarted, I think from when the cluster was updated). So presumably it succeeded in getting the graph afterwards (would need to increase verbosity to be able to tell I think). Is your node updated to the latest release?
Manually curling the server seems to work fine right now at least:
$ curl -L 'https://updates.coreos.fedoraproject.org/v1/graph?basearch=x86_64&stream=testing&rollout_wariness=0'
{
"nodes": [
{
"version": "30.20190716.1",
"metadata": {
"org.fedoraproject.coreos.releases.age_index": "0",
"org.fedoraproject.coreos.updates.deadend": "true",
"org.fedoraproject.coreos.scheme": "checksum",
"org.fedoraproject.coreos.updates.deadend_reason": "https://github.com/coreos/fedora-coreos-tracker/issues/215"
},
"payload": "ff4803b069b5a10e5bee2f6bb0027117637559d813c2016e27d57b309dd09d6f"
},
...
I'm on the last release and restarting Zincati / status tells everything is OK.
Looks like a false alarm. Sorry for the noise.
Since January 24th, 2024 on my server: