← Back to team overview

yahoo-eng-team team mailing list archive

[Bug 1628017] Re: unable to access vm by floating ip from vm without floating

 

Ok, so "Connection refused" was a result of stale ip address on rfp
device that was not deleted after l3 agent restart with new code. If
recreate instances/floating ip from scratch everything works fine. I'm
going to backport the fix to stable/mitaka. Marking this as invalid
since the problem should be fixed on master.

** Changed in: neutron
       Status: New => Invalid

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1628017

Title:
  unable to access vm by floating ip from vm without floating

Status in neutron:
  Invalid

Bug description:
  Steps to reproduce:

  1. create 2 machines in one internal network. Make sure that VMs created on one compute node
  2. assign floating ip to one vm
  3. try to connect from second vm (without floating ip) to vm with floating ip
  like nc -v floating_ip 22

  Expected result
  ===============
  nc -v 192.168.120.116 22
  Connection to 192.168.120.116 22 port [tcp/ssh] succeeded!
  SSH-2.0-OpenSSH_6.6.1p1 Ubuntu-2ubuntu2.7

  Actual result
  =============
  nc -v 192.168.120.116 22
  nc: connect to 192.168.120.116 port 22 (tcp) failed: Connection timed out


  BTW, if I ping floating ip - I get internal icmp response
  ping 192.168.120.116
  PING 192.168.120.116 (192.168.120.116) 56(84) bytes of data.
  64 bytes from 192.168.111.20: icmp_seq=1 ttl=64 time=0.513 ms
  64 bytes from 192.168.111.20: icmp_seq=2 ttl=64 time=0.538 ms

  
  This bug could be reproduced only if VMs created on the same compute node, if I migrate one VM to another - I am able to access floating ip

  Environment
  ===========
  1. Mirantis Openstack 9.0.1 with DVR enabled

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/1628017/+subscriptions


References