[OmniOS-discuss] Zone "lost" its IP, was then bound to global zone

[Bryan Horstmann-Allen]

The global zone:

# grep network /etc/zones/icg_mx.xml
   <network address="a.b.c.96" physical="igb1"/>

igb1:16: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 2
        inet a.b.c.96 netmask ffffff00 broadcast 208.72.237.255

Note the lack of a "zone: icg_mx" field in that alias output.

The zone:

# ifconfig -a
lo0:14: flags=2001000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4,VIRTUAL> mtu 8232 index 1
        inet 127.0.0.1 netmask ff000000
lo0:14: flags=2002000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv6,VIRTUAL> mtu 8252 index 1
        inet6 ::1/128

Any connections to a.b.c.96 went to the global zone instead of the zone itself.

How I "fixed" it:

  # ifconfig igb1:16 down
  # ifconfig igb1:16 unplumb
  # zoneadm -z icg_mx reboot

In 7 years of heavy zone usage, I've never seen anything like this.

This caused about 11 hours of internal automated mail (systems, sev2
monitoring, cron, etc) -- thankfully just that -- to bounce, since it was all
going to the gz, instead of the zone.

To put it _very_ mildly, this is _really_ concerning.  No humans were logged
into the machine when this occurred. The config mgmt daemon was disabled. I'm
going by when mail stopped being accepts on the ngz; which was still running...
just missing its IP.

No useful logs to syslog, SMF, etc.

Has anyone run into this?
-- 
bdha
cyberpunk is dead. long live cyberpunk.