Digging somewhat deeper, I think this problem might be related to a bad interaction between the kernel, systemd, and docker's network driver. I built an Alma9 container image using the xrd-docker script adapted to work with podman instead of docker and the tests did run fine in that case (apart from commit eae57b4, which causes a deadlock in the PostMaster tests and has been locally reverted).

Just for the record, I see some output in the system logs which are likely related.

On the host:

May 24 15:08:39 epsftws.cern.ch dhcpcd[746]: veth711de2d: deleting address fe80::394f:a6d:e504:6a66
May 24 15:08:39 epsftws.cern.ch dhcpcd[746]: veth711de2d: deleting route to 169.254.0.0/16
May 24 15:08:39 epsftws.cern.ch (udev-worker)[383066]: Network interface NamePolicy= disabled on kernel command line.
May 24 15:08:39 epsftws.cern.ch kernel: docker0: port 1(veth711de2d) entered disabled state
May 24 15:08:39 epsftws.cern.ch kernel: veth711de2d (unregistering): left allmulticast mode
May 24 15:08:39 epsftws.cern.ch kernel: veth711de2d (unregistering): left promiscuous mode
May 24 15:08:39 epsftws.cern.ch kernel: docker0: port 1(veth711de2d) entered disabled state
May 24 15:08:39 epsftws.cern.ch dhcpcd[746]: veth711de2d: removing interface
May 24 15:08:40 epsftws.cern.ch systemd[1]: run-docker-netns-0473608522e4.mount: Deactivated successfully.

and inside the metaman container:

May 24 13:08:53 metaman kernel: docker0: port 1(veth711de2d) entered blocking state
May 24 13:08:53 metaman kernel: docker0: port 1(veth711de2d) entered disabled state
May 24 13:08:53 metaman kernel: veth711de2d: entered allmulticast mode
May 24 13:08:53 metaman kernel: veth711de2d: entered promiscuous mode
May 24 13:08:53 metaman kernel: docker0: port 1(veth711de2d) entered blocking state
May 24 13:08:53 metaman kernel: docker0: port 1(veth711de2d) entered forwarding state
May 24 13:08:53 metaman kernel: docker0: port 1(veth711de2d) entered disabled state
May 24 13:08:53 metaman kernel: IPv6: ADDRCONF(NETDEV_CHANGE): veth711de2d: link becomes ready
May 24 13:08:53 metaman kernel: docker0: port 1(veth711de2d) entered blocking state
May 24 13:08:53 metaman kernel: docker0: port 1(veth711de2d) entered forwarding state
May 24 13:08:53 metaman kernel: docker0: port 1(veth711de2d) entered disabled state
May 24 13:08:53 metaman kernel: eth0: renamed from vethd46c78b
May 24 13:08:53 metaman kernel: docker0: port 1(veth711de2d) entered blocking state
May 24 13:08:53 metaman kernel: docker0: port 1(veth711de2d) entered forwarding state
May 24 13:08:53 metaman kernel: docker0: port 1(veth711de2d) entered disabled state
May 24 13:08:53 metaman kernel: vethd46c78b: renamed from eth0
May 24 13:08:53 metaman kernel: docker0: port 1(veth711de2d) entered disabled state
May 24 13:08:53 metaman kernel: veth711de2d (unregistering): left allmulticast mode
May 24 13:08:53 metaman kernel: veth711de2d (unregistering): left promiscuous mode
May 24 13:08:53 metaman kernel: docker0: port 1(veth711de2d) entered disabled state

So this is a networking problem with docker, not a problem with XRootD.

—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.

[ { "@context": "http://schema.org", "@type": "EmailMessage", "potentialAction": { "@type": "ViewAction", "target": "https://github.com/xrootd/xrootd/issues/1990#issuecomment-1561937370", "url": "https://github.com/xrootd/xrootd/issues/1990#issuecomment-1561937370", "name": "View Issue" }, "description": "View this Issue on GitHub", "publisher": { "@type": "Organization", "name": "GitHub", "url": "https://github.com" } } ]

Use REPLY-ALL to reply to list

To unsubscribe from the XROOTD-DEV list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-DEV&A=1