docker-archive / classicswarm

Swarm Classic: a container clustering system. Not to be confused with Docker Swarm which is at https://github.com/docker/swarmkit
Apache License 2.0
5.75k stars 1.08k forks source link

The swarm node disappeared after node successfully has joined a few minutes #2824

Closed nighteblis closed 6 years ago

nighteblis commented 6 years ago

The node disappeared after node successfully has joined a few minutes.

$ sudo docker --version Docker version 17.09.0-ce, build afdb6d4

At the worker node , sudo docker swarm join . I could see the node ready at the manager node with the command "sudo docker node ls ". But after one minute, the node disappeared. BTW, The manager and the worker node are not in the same Ethernet Subnet, They could be connected with each other.

Steps: master: sudo docker swarm init (172.16.2.x) slave: sudo docker swarm join ... (10.10.x.x)

master and slave are not in same subnet , but could ping each other. Both node have no iptables rules.

The error logs on slave node.

level=error msg="agent: session failed" error="rpc error: code = 7 desc = Permission denied: unauthorized peer role: rpc error: code = PermissionDenied desc = Permission denied: node 44whub2gp4vhkmf5vyuuh571y was removed from swarm" module="node/agent"

lh83mail commented 6 years ago

i got the same problem, any idea?

tail /var/log/syslog

Dec 5 14:15:08 gcsvr03 dockerd[32193]: time="2017-12-05T14:15:08.068140152+08:00" level=error msg="agent: session failed" error="rpc error: code = 7 desc = Permission denied: unauthorized peer role: rpc error: code = PermissionDenied desc = Permission denied: node 44whub2gp4vhkmf5vyuuh571y was removed from swarm" module="node/agent" Dec 5 14:15:14 gcsvr03 dockerd[32193]: time="2017-12-05T14:15:14.375344762+08:00" level=error msg="agent: session failed" error="rpc error: code = 7 desc = Permission denied: unauthorized peer role: rpc error: code = PermissionDenied desc = Permission denied: node 44whub2gp4vhkmf5vyuuh571y was removed from swarm" module="node/agent" Dec 5 14:15:16 gcsvr03 dockerd[32193]: time="2017-12-05T14:15:16.939711544+08:00" level=error msg="agent: session failed" error="rpc error: code = 7 desc = Permission denied: unauthorized peer role: rpc error: code = PermissionDenied desc = Permission denied: node 44whub2gp4vhkmf5vyuuh571y was removed from swarm" module="node/agent"

nighteblis commented 6 years ago

@lh83mail Refered to your logs , this may be a permission problems. However I don't know why docker engine told me "already joined" when I try to re-join swarm node. That means the node was not removed actually ?

nishanttotla commented 6 years ago

@nighteblis please open this issue on https://www.github.com/docker/swarmkit. Sorry for the trouble.

nighteblis commented 6 years ago

@nishanttotla you mean it's the kit problem? Do you know why the permission error happend ?
BTW Opend a new issue on swarmkit#2612