I've been seeing yo-yo'ing connectivity checks from NetworkManager on Fedora systems for a number of years now, and it seems that this could be due to one flaky host in its round-robin.
In this ticket, I got debug from NetworkManager about the connectivity checks it was making, and to whom: https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/issues/648
And it seems that more often than not, calls to 85.236.55.6 will timeout. Is this known/expected?
85.236.55.6
Should NetworkManager's connectivity check use a separate subdomain where only servers with very high availability are listed?
We are currently hitting an issue where some of our proxies (mostly the older/slower ones) stop processing new connections. :(
It's under investigation. I don't think it's particular to that proxy, it's happening to a number of them.
Sorry for the issue. ;(
@phsmoura is digging more into this, so far we have not been able to fully pin down a cause. ;(
Metadata Update from @kevin: - Issue priority set to: Waiting on Assignee (was: Needs Review) - Issue tagged with: high-gain, high-trouble, ops
We believe there is a bug in Apache causing child processes to fail to start here. We are rolling back httpd to 2.4.51 on proxy09 & proxy 31 to see if this resolves the issue.
Metadata Update from @aheath1992: - Issue close_status updated to: Fixed with Explanation - Issue status updated to: Closed (was: Open)
After some research we determined we were hitting Apache bug 65769. We have downgraded the proxies from httpd-2.4.52-1 to httpd-2.4.51-1 which has stabilized the environment. We will upgrade once a new stable version is out that is not affect by the bug.
Login to comment on this ticket.