node preseed retrieval fails due to euca_find_cluster returning

Bug #445738 reported by Dustin Kirkland  on 2009-10-07
This bug affects 2 people
Affects Status Importance Assigned to Milestone
eucalyptus (Ubuntu)
Dustin Kirkland 

Bug Description

The cluster controller's ip information looks like this:

2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:16:d3:ca:e8:1e brd ff:ff:ff:ff:ff:ff
    inet brd scope global eth0
    inet scope global eth0
    inet6 fe80::216:d3ff:feca:e81e/64 scope link
       valid_lft forever preferred_lft forever

When avahi is asked for this system's ip address by its avahi hostname, sometimes it returns, and sometimes it returns

Unfortunately, isn't very useful to other nodes.

We already have some special case logic in euca_find_cluster that prefers an ipv4 address over an ipv6 address.

We need some additional logic to avoid the, and use the suggested ip address of the CC.

By "suggested IP address", a secondary change is necessary in the avahi-publish to use the $CC_IP_ADDR of the system as the "name" of the service. Previously, this was $(hostname).


Changed in eucalyptus (Ubuntu):
status: New → In Progress
importance: Undecided → High
assignee: nobody → Dustin Kirkland (kirkland)
Matt Zimmerman (mdz) wrote :

> inet scope global eth0

Shouldn't that be using "scope link"? I don't know if that would matter for this bug, but it seems a bit odd for it to be a global scope address. Might be harmless though.

Changed in eucalyptus (Ubuntu):
status: In Progress → Fix Committed
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package eucalyptus - 1.6~bzr916-0ubuntu1

eucalyptus (1.6~bzr916-0ubuntu1) karmic; urgency=low

  * Merge upstream revision 916
  * This snapshot is expected to fix the following bugs:
    - LP: #443314 - fix system address selection
    - LP: #443325 - fix regression in heartbeat, fix NPE in system mode
    - LP: #444838 - fix meta-data service in multi-cluster
    - LP: #444908 - fix running instances on multi-cluster
    - LP: #445206 - fix remote bootstrap on walrus+sc
  * debian/eucalyptus-cc.eucalyptus-cc-publication.upstart,
    debian/local/euca_find_cluster.c: when doing the avahi-publish, use
    $CC_IP_ADDR rather than $(hostname) for the "name" of the service;
    use this ip address in the case where euca_find_cluster gets the
    link-local address, LP: #445738

 -- Dustin Kirkland <email address hidden> Wed, 07 Oct 2009 20:01:56 -0500

Changed in eucalyptus (Ubuntu):
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers