I have installed a two-node sun cluster 3.2 x86 SO Oracle Solaris 10 u09 .
My problem is that I installed a non-global zone in the first node and when a try to boot that zone the cluster is restarted.
I find this log
[ID 562397 kern.notice] Failfast: Aborting zone "global" (zone ID 0) because "sc_zonesd" died 35 seconds ago.
[ID 570001 auth.error] reboot after panic: Failfast: Aborting zone "global" (zone ID 0) because "sc_zonesd" died 35 seconds ago.
Could please anyone help me
Thanks in advance
Several critical processes within Solaris Cluster register with the failfast daemon and I guess this is one of them. If the registered process fails, then the failfast daemon will panic the node to preserve cluster integrity. That's the explanation of the panic, not that it tells you why it happened.
How did you create the zone? Are there any messages in the messages file for the zone? You can get them through the zonepath/root/var/adm/messages. Are there any other messages in the global zone /var/adm/messages file that might indicate an issues further back in the log?
Thanks for your answer
I created a very very basic zone like that
zonecfg -z test-zone
zona1: No such zone configured
Use 'create' to begin configuring a new zone.
zonecfg:test-zone> remove inherit-pkg-dir dir=/sbin
zonecfg:test-zone> remove inherit-pkg-dir dir=/usr
zonecfg:test-zone> remove inherit-pkg-dir dir=/platform
zonecfg:test-zone> remove inherit-pkg-dir dir=/lib
zonecfg:test-zone> set autoboot=false
zonecfg:test-zone> set zonepath=/zones/test-zone
zonecfg:test-zone> add net
zonecfg:test-zone:net> set physical=e1000g0
zonecfg:test-zone:net> set address=192.168.31.136
zoneadm -z test-zone install
zoneadm -z test-zone boot ---> In this point the cluster node crash
There is no logs information about that in the messages file. When I doesn´t boot the zone the cluster node is up and running.
Tim , do you know if there is some procedure to create a zone in a cluster node with sun cluster, I will appreciate.
There is an example of creating a zone for a cluster on page 379 of my book "Oracle Solaris Cluster Essentials". However, apart from the removal of the inherited pkgs, it is very similar. The only difference is that I qualified my address with /24.
Why don't you try the procedure without the remove statements first and see if that works on your system? Then move on to a whole-root zone. I'm wondering if one of the 'removes' should not be there (or causes a bug).