eucalyptus logs

Submitted by admin on Tue, 2008-08-12 10:14.

> Hello Fabio,
>
> thanks for using Eucalyptus! First, to rule out some possible transients issues, can you confirm that you restarted Eucalyptus (/etc/init.d/eucalyptus restart) on both the frontend and the containers? Upon restart it may takes few minutes for ec2-describe-availability-zones to report correctly.

Hello Graziano, thanks for your fast reply, I tried many times to restart the whole cluster and too reboot too al the nodes and the frontend but nothing changed.

>
> Then can you check what's in /opt/eucalyptus/var/log/eucalyptus and perhaps send it to us? In particular the cc.log (on the frontend) and the nc.log (on the compute nodes) may have some information on why they are not talking to each other. Perhaps the keys aren't in sync: in this case you can try to re-push them to the node (using /opt/eucalyptus/usr/sbin/euca_sync_key).

I've re-pushed keys, also adding the flag -x to the euca_sync_key script to be sure it is working, here follows the output:

[root@cluster eucalyptus-1.1]# ./usr/sbin/euca_sync_key
+ EUCALYPTUS=
+ KEYS=
+ CONF=
+ FORCE=N
+ VERBOSE=Y
+ '[' -e /etc/default/eucalyptus ']'
+ CONF=/etc/default/eucalyptus
+ '[' 0 -gt 0 ']'
+ '[' -z /etc/default/eucalyptus -o '!' -e /etc/default/eucalyptus ']'
+ echo 'Using /etc/default/eucalyptus as configuration'
Using /etc/default/eucalyptus as configuration
+ . /etc/default/eucalyptus
++ EUCALYPTUS=/opt/eucalyptus-1.1
++ START_CLOUD=Y
++ START_CC=Y
++ START_NC=N
++ ENABLE_WS_SECURITY=Y
++ CLOUD_PORT=8773
++ CLOUD_SSL_PORT=8443
++ CC_PORT=8774
++ NODES='vm-container-0-0 vm-container-0-1'
++ NC_SERVICE=axis2/services/EucalyptusNC
++ NC_PORT=8774
++ TEMPLATE_PATH=/share/apps/eucalyptus/templates
++ REGISTERED_PATH=/share/apps/eucalyptus/registered
++ INSTANCE_PATH=/usr/local/eucalyptus/instances
++ VNET_PORT=1978
++ VNET_DHCPDEAMON=/usr/sbin/dhcpd
++ VNET_PUBINTERFACE=eth1
+ '[' -z /opt/eucalyptus-1.1 ']'
+ '[' N '!=' Y ']'
+ '[' '!' -x /opt/rocks/bin/rocks ']'
+ '[' -z 'vm-container-0-0 vm-container-0-1' ']'
++ echo vm-container-0-0 vm-container-0-1
+ for x in '`echo $NODES`'
+ '[' Y = Y ']'
+ rsync -avz -e ssh /opt/eucalyptus-1.1/var/eucalyptus/keys vm-container-0-0:/opt/eucalyptus-1.1/var/eucalyptus
+ for x in '`echo $NODES`'
+ '[' Y = Y ']'
+ rsync -avz -e ssh /opt/eucalyptus-1.1/var/eucalyptus/keys vm-container-0-1:/opt/eucalyptus-1.1/var/eucalyptus
[root@cluster eucalyptus-1.1]#

(I used eth1 as the vnet public interface, eth1 is the external nic of my cluster on the 192.168.1.0 net, and I'm using my notebook on the same 192.168.1.0 net to launch ec2* commands)

These are the cc.log on the frontend:

[Tue Aug 12 12:01:10 2008] DescribeInstances(): done
[Tue Aug 12 12:01:25 2008] DescribeResources(): called 1
[Tue Aug 12 12:01:25 2008] vnetStartNetwork(): start_vdeswitch(/opt/eucalyptus-1.1/var/run/eucalyptus/net,clceuca0) == 0
[Tue Aug 12 12:01:25 2008] vnetStartNetwork(): configure_vdeswitch(/opt/eucalyptus-1.1/var/run/eucalyptus/net,clceuca0,-1) == 0
[Tue Aug 12 12:01:25 2008] vnetStartNetwork(): start_port(/opt/eucalyptus-1.1/var/run/eucalyptus/net,clceuca0) == 0
[Tue Aug 12 12:01:25 2008]      Running cmd: vde_cryptcab -s /opt/eucalyptus-1.1/var/run/eucalyptus/net/clceuca0.ctl -c cluster.bee-side.com:1978 -P /opt/eucalyptus-1.1/var/eucalyptu
s/keys/vdekey -p 29313
[Tue Aug 12 12:01:25 2008] vnetStartNetwork():  check_wire(), wirepids: 9284, host: cluster.bee-side.com (inactive)
[Tue Aug 12 12:01:25 2008] vnetStartNetwork(): done
[Tue Aug 12 12:01:25 2008] ncStub call done: 9285 0
[Tue Aug 12 12:01:25 2008]      -417/0 4749/4749 1/1
[Tue Aug 12 12:01:26 2008] ncStub call done: 9286 0
[Tue Aug 12 12:01:26 2008]      -33/0 4749/4749 1/1
[Tue Aug 12 12:01:26 2008] refresh_resources(): done
[Tue Aug 12 12:01:26 2008] DescribeResources(): done
[Tue Aug 12 12:01:26 2008] RegisterImage(): called
[Tue Aug 12 12:01:26 2008]      input params: location ttylinux, userId eucalyptus, amiId emi-0D05022C
[Tue Aug 12 12:01:26 2008] scRegisterImage(): user, amiId, location: eucalyptus emi-0D05022C ttylinux
[Tue Aug 12 12:01:26 2008]      template_path: /share/apps/eucalyptus/templates/ttylinux
[Tue Aug 12 12:01:26 2008] RegisterImage(): done
[Tue Aug 12 12:01:26 2008]      reg_path2: /share/apps/eucalyptus/registered/eucalyptus/emi-0D05022C
[Tue Aug 12 12:01:26 2008]      rc from state: 0
[Tue Aug 12 12:01:26 2008] StartNetwork(): called
[Tue Aug 12 12:01:26 2008] vnetStartNetwork(): start_vdeswitch(/opt/eucalyptus-1.1/var/run/eucalyptus/net,clceuca0) == 0
[Tue Aug 12 12:01:26 2008] vnetStartNetwork(): configure_vdeswitch(/opt/eucalyptus-1.1/var/run/eucalyptus/net,clceuca0,-1) == 0
[Tue Aug 12 12:01:26 2008] vnetStartNetwork(): start_port(/opt/eucalyptus-1.1/var/run/eucalyptus/net,clceuca0) == 0
[Tue Aug 12 12:01:26 2008]      Running cmd: vde_cryptcab -s /opt/eucalyptus-1.1/var/run/eucalyptus/net/clceuca0.ctl -c cluster.bee-side.com:1978 -P /opt/eucalyptus-1.1/var/eucalyptu
s/keys/vdekey -p 29313
[Tue Aug 12 12:01:26 2008] vnetStartNetwork():  check_wire(), wirepids: 9296, host: cluster.bee-side.com (inactive)
[Tue Aug 12 12:01:26 2008] vnetStartNetwork(): done
[Tue Aug 12 12:01:26 2008] StartNetwork(): SUCCESS return from vnetStartNetwork 0
[Tue Aug 12 12:01:26 2008] StartNetwork(): done

and nc.log on one node:
[Tue Aug 12 12:01:25 2008] doDescribeResource() invoked
[Tue Aug 12 12:01:26 2008] StartNetwork(): called
[Tue Aug 12 12:01:26 2008] vnetStartNetwork(): start_vdeswitch(/opt/eucalyptus-1.1/var/run/eucalyptus/net,nceuca0) == 0
[Tue Aug 12 12:01:26 2008] vnetStartNetwork(): configure_vdeswitch(/opt/eucalyptus-1.1/var/run/eucalyptus/net,nceuca0,-1) == 0
[Tue Aug 12 12:01:26 2008] vnetStartNetwork(): start_port(/opt/eucalyptus-1.1/var/run/eucalyptus/net,nceuca0) == 0
[Tue Aug 12 12:01:26 2008] vnetStartNetwork(): done
[Tue Aug 12 12:01:26 2008] StartNetwork(): SUCCESS return from vnetStartNetwork 0
[Tue Aug 12 12:01:26 2008] StartNetwork(): done
[Tue Aug 12 12:01:42 2008] doDescribeInstances() invoked
[Tue Aug 12 12:01:57 2008] doDescribeResource() invoked
[Tue Aug 12 12:02:12 2008] doDescribeInstances() invoked
[Tue Aug 12 12:02:28 2008] doDescribeResource() invoked

Looking at the cc.log the cluster.bee-side.com is described as "inactive" but I don't know how to change that status maybe something around here?

>
> Finally, Eucalyptus uses the availability zones as a way to distinguish different cluster controller (and clusters associated with it). So in your case you should be able to see only one availability zone the one related to your fronted (cluster.bee-side.com). Once the nodes will register correctly with the frontend, you should be able to use them all without having to specify any zone at all. The error you see now happens because Eucalyptus doesn't see any available node (the 000/000 you see means 0 nodes of 0 are available) and it should be cleared out once the nodes reports correctly.

Ok thanks also for this clarification, so 1 physical node equals to 1 "resource" in general? What I'm asking is: does a physical machine host only one ami at time?

Thanks in advance
All my bests
Fabio
>
> graziano