Actions
Bug #4491
closedRudder considers an inventory as "failed" if the endpoint is unavailable
Status:
Released
Priority:
1 (highest)
Assignee:
Category:
Web - Nodes & inventories
Target version:
Pull Request:
Severity:
UX impact:
User visibility:
Effort required:
Priority:
Name check:
Fix check:
Regression:
Description
Inventories processed on the Rudder root server are stored in a WebDAV directory (/var/rudder/inventories/incoming) then periodically sent to the endpoint on the server.
The endpoint runs a pre-check, and if the file is invalid, returns a 412 HTTP error code. We use that to move the invalid inventory to /var/rudder/inventories/failed. This was introduced in #3139.
However, if the endpoint is not available (eg, jetty has not yet started), then we also, naively, move the files (that may be correct) to /var/rudder/inventories/failed. This means that we can lose incoming inventories.
This should not be the case.
Actions