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
seaweedfs 2.49 - weed server started with IPv4 address only listens on IPv6 #2098
Comments
use |
The older -ip flag is now just a way to define an instance/hostname string?
Yes, -ip.bind works, although I couldn't figure out how to bind to * on ipv4 for further tests. Something still goes wrong:
Seems like the volume server tries to connect to 192.168.1.130:19333 before master is running and obviously fails. But even after it is started (apart from seeing the LISTENing ports, I can reach the web interface at 192.168.1.130:9333, with no volumes) there is still the repeating error at the end of the log. Does it actually give up on retrying gRPC on IPv4 after the first failure or such? |
Almost the same error on the 2.50 release when using explicit ip.bind to 192.168.1.130 as in the last comment. It now just alternates between two different error messages with an new "use of closed network connection":
|
The logs showed |
That was before the additional error which now happens in 2.50. What runs is currently:
I deleted the older logs prior to restarting to be certain this can only be from the current attempt:
|
Maybe try to make it reproducible via some docker instance first. |
I can check if it works in Podman (no Docker on that machine, unfortunately). Seems like the default compose setup is running the parts individually rather than all in one - I'll try that. |
Figured it out thanks to trying more with the containers. These remaining issues were not due to IPv4/IPv6 networking but the security certificates coincidentially had additionally expired just now (they have a short expiration time based on certstrap defaults). Results in the above errors and weed shell hanging rather than more obvious verbose complaints about expired certificates, apparently. Thanks for the help with this issue. |
Thanks for the update! |
version 8000GB 2.49 42fb03a linux amd64
I invoke weed server like this (via a systemd service that runs it under a "seaweedfs" user):
weed server -rack=aaa -ip=192.168.1.130 -master.defaultReplication=010 -master.peers=192.168.1.130:9333 -filer=true -filer.port=8515 -filer.defaultReplicaPlacement=010 -volume.port=8526 -volume.max=56 -dir=/mnt/seaweedfs
192.168.1.130 is also that same machine's IPv4. But this happens:
I observe on "lsof -i -P -n | grep LISTEN" that weed seems to have started gRPC on port + 10000, but only listening on IPv6. No mention of IPv4:
I imagine this is the problem, but I can't figure out what made seaweedfs use IPv6 when an IPv4 address was specified. Nothing was using 9333 or 19333 on IPv4. Perhaps it is a bug?
The text was updated successfully, but these errors were encountered: