Replies: 3 comments 4 replies
-
Reboot would be easiest, if not stop all containers, then remove the stale netns-* files and remove the ipam.db from |
Beta Was this translation helpful? Give feedback.
-
A workaround that I've been able to use successfully: Create a dummy network that uses the exhausted IP range and let your containers create a new one.
|
Beta Was this translation helpful? Give feedback.
-
Since a month or two I'm also facing this issue on one of my servers. I'm unable to track the source down. Did you find out what was triggering it for you? For me it happens on one of my Fedora CoreOS servers, but not on the other one. All few weeks all my services fail to start (I guess after an update or so), forcing me to manually restart the server. |
Beta Was this translation helpful? Give feedback.
-
I got into a situation that sounds a lot like #18615. For some reason, maybe because of a continuously restarting container, I seem to have run out of IP addresses:
(I'm using podman-compose to start paperless-ngx.)
ls -l /run/user/1000/netns/netns-*
shows 257 empty files named likenetns-ff176f26-1f2e-edc4-3417-2425046f394c
with different IDs.Now, the GitHub issue I've linked above mentions that this has been fixed in Podman 4.8, but is there a way for me to fix this manually without rebooting the machine? I can't really update to 4.8 right now.
podman network rm paperless_default
returns successfully, removes the network, but thenetns
files remain, and so does thefailed to find free IP
error.ip netns list
produces no output.How do I fix this? Any help greatly appreciated.
Beta Was this translation helpful? Give feedback.
All reactions