-
Notifications
You must be signed in to change notification settings - Fork 180
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Logout sometimes fails with 401 response #160
Comments
I'm guessing the BMC you're using might not be properly closing sessions. Would it be possible to enable the tracing facilities in the library and reproduce the issue? The following code block should create a log file for your with all HTTP traces captured to the file "RedfishHTTPDebug.log".
|
I'll try to add this capability in our product to try to put some light in this issue. We identify that this is happening on Dell and HPE servers: BIOS Version: 2.21.2 BIOS Version: 2.19.1 I also tried to stress one server by opening multiple sessions (20 sessions) and continuously making requests. It also failed with 401, but to create a new session and not to delete as the bug description.
|
Redfish sometimes fails to logout after scraping the data and this end up generating the metric redfish_collector_failed that trigger prometheus alerts that can be noise. This can be a BMC problem that might not be properly closing sessions. For more info see: DMTF/python-redfish-library#160 Closes: canonical#76
If you're seeing it from different firmware on different vendors, then that makes it less likely a BMC issue. Perhaps the library is hitting a timeout or other networking issue when it issues a DELETE to the session, which would invoke a retry, and if the BMC already processed the DELETE, a 401 would return... For that stress test you tried, it's possible you might be hitting a session limit on the BMC. At least the systems I have will return a 503 once you hit their session limit, so it might be something else entirely different. I'll double check, but the logging traces I recommended adding should capture retry attempts as well. |
* Igonore if redfish fails to logout with 401 as as response Redfish sometimes fails to logout after scraping the data and this end up generating the metric redfish_collector_failed that trigger prometheus alerts that can be noise. This can be a BMC problem that might not be properly closing sessions. For more info see: DMTF/python-redfish-library#160 Closes: #76
I was not able to manually reproduce it, but sometimes it seems that the logout is failing with 401.
From the logs it's possible to create a session and fetch information, but if fails to logout. I'm not understanding how the authorization can break during this process.
originally found on canonical/prometheus-hardware-exporter#76
The text was updated successfully, but these errors were encountered: