Hi there. How do I troubleshoot the non-responsiveness of docker commands, or the inability to access the UI on port 8080?
My system seemed to be working ok, but then my UI stopped responding. (This site can’t be reached…refused to connect…ERR_CONNECTION_REFUSED).
Through the console, I can log in to RancherOS, and can ping the outside world, and ros config seems to respond (tested with ros config export).
But ‘docker container ls’ fails to return a response…the console just waits and waits, and eventually Ctrl+C returns me to the prompt to enter a new command.
Please help. How do I diagnose this?
PS. might be worth mentioning that I changed my config to use a static ip as indicated here, and did a “sudo reboot” for the change to take effect. But after that reboot all was still working ok. But if I did break it, how do I REMOVE that ‘network’ section from my config, which is shown in this image:
Not being able to access the UI is usually due to wrong MTU settings when running on a cloud provider. You explicitly set 1500 in the config, I assume that’s correct for where you are running?
docker container ls
being stuck is usually not because of that (possibly DNS but don’t think so). You can check sudo system-docker logs docker
, sudo cat /var/log/docker.log
. If this doesn’t show anything, you can run Docker in debug and see what it reports.
mmm I must sheepishly admit that I set mtu based on what I saw in the referenced documentation. How can I remove that line from my config to see if it resolves the problem? In fact, how can I remove the entire ‘network’ section from that config?
Ok it turned out this is as simple as editing the /var/lib/rancher/conf/cloud-config.yml file. I removed the mtu setting line completely, and rebooted.
My UI is back, and the docker commands respond again.
Now I just got to figure out why I can’t access my crashplan-pro UI on port 5800, even though the container is listed as running.
Ok I’m feeling quite sad (frustrated) now.
After getting my UI up and running, I noticed that I cannot access my crashplan-pro container’s UI (on port 5800).
I then checked out the Hosts view, and it shows the host as Disconnected.
Back to this forum, and I found this thread which did not help, but the opening post mentions that a restart of his FreeNAS VM eventually reconnects the Host.
So I restarted my FreeNAS VM, and now I am once again unable to access my Rancher UI on port 8080, and the docker commands once again just wait and wait forever!!!
A ros config export shows that the network interface is still as per my last edit (no custom mtu setting there).
What on earth is going on here?
Ok, so I did a
sudo reboot
and now my docker commands respond, and my UI is accessible again on port 8080.
Still my Host shows as DISCONNECTED, the crashplan-pro container on that host shows as running, but I can’t access it’s UI on port 5800. Sigh. Will open/contribute to another thread for that issue.