Created attachment 704743 [details]
nc.log

Description of problem:
The eucalyptus-nc is suppose to send the number of available resources back to the cluster controller, but it does not correctly send that back when an instance is created on the NC

Version-Release number of selected component (if applicable):
eucalyptus-nc-3.2.0-0.5.20121210git5ab72572.fc18.x86_64


How reproducible:
The steps to reproduce this problem are given below

Steps to Reproduce:
1. Setup a small Eucalyptus cloud with 2 machines, machine 1 acting as the front-end having the CLC/Walrus/CC/SC and machine 2 running NC
2. Configure the networking mode as MANAGED-NOVLAN 
3. Configure SELinux to Disabled
4. Upload a EMI to your Eucalyptus cloud and run an instance off the EMI
5. Keep checking the /var/log/eucalyptus/nc.log for the function doDescribeResource, which tells the resource (cpu/memory/disk) stats


Actual results:
Pay attention to the timestamps:

2013-03-03 18:27:10  INFO 000013499 startup_thread           | [i-F1D9426B] booting
2013-03-03 18:27:10 DEBUG 000013499 change_state             | [i-F1D9426B] state change for instance: Staging -> Booting (Extant)
2013-03-03 18:27:11 DEBUG 000013496 change_state             | [i-F1D9426B] state change for instance: Booting -> Running (Extant)
2013-03-03 18:27:12 DEBUG 000013556 isConfigModified         | current mtime=1362311919, stored mtime=0
2013-03-03 18:27:13 DEBUG 000013376 doDescribeResource       | returning cores=4/4 mem=7118/7118 disk=84/84 iqn=iqn.1994-05.com.redhat:df94a7c9a460
2013-03-03 18:27:13 DEBUG 000013386 doDescribeInstances      | invoked userId=eucalyptus correlationId=UNSET epoch=0 services[0]{.name= .type= .uris[0]=}
2013-03-03 18:27:19 DEBUG 000013393 doDescribeResource       | returning cores=4/4 mem=7118/7118 disk=84/84 iqn=iqn.1994-05.com.redhat:df94a7c9a460
2013-03-03 18:27:20 DEBUG 000013424 doDescribeInstances      | invoked userId=eucalyptus correlationId=UNSET epoch=0 services[0]{.name= .type= 
2013-03-03 18:27:20 DEBUG 000013428 doAssignAddress          | [i-F1D9426B] invoked (publicIp=10.104.3.50)
2013-03-03 18:27:26 DEBUG 000013433 doDescribeResource       | returning cores=2/4 mem=6094/7118 disk=69/84 iqn=iqn.1994-05.com.redhat:df94a7c9a460
2013-03-03 18:27:26 DEBUG 000013401 doDescribeInstances      | invoked userId=eucalyptus correlationId=UNSET epoch=0 services[0]{.name= .type= .uris[0]=}
2013-03-03 18:27:32 DEBUG 000013406 doDescribeResource       | returning cores=4/4 mem=7118/7118 disk=84/84 iqn=iqn.1994-05.com.redhat:df94a7c9a460
2013-03-03 18:27:32 DEBUG 000013370 doDescribeInstances      | invoked userId=eucalyptus correlationId=UNSET epoch=0 services[0]{.name= .type= .uris[0]=}
2013-03-03 18:27:32 DEBUG 000013370 doDescribeInstances      | [i-F1D9426B] Extant pub=10.104.3.50 priv=192.168.191.166 mac=D0:0D:F1:D9:42:6B vlan=768 net=38 plat=linux vols=
2013-03-03 18:27:38 DEBUG 000013367 doDescribeResource       | returning cores=2/4 mem=6094/7118 disk=69/84 iqn=iqn.1994-05.com.redhat:df94a7c9a460
2013-03-03 18:27:38 DEBUG 000013366 doDescribeInstances      | invoked userId=eucalyptus correlationId=UNSET epoch=0 services[0]{.name= .type= .uris[0]=}

Expected results:
It is expected that as soon as new instance creation request arrives at the NC, it should update the resource availability, so correct information is passed up the layer and it should maintain that during the complete lifecycle of the instance, instead of changing the resource statistics every time (during the instance lifecycle) randomly. 

Additional info:
nc.log from the system showing this symptom is attached to this ticket.