sysdrain-offline.sh:
USAGE: ./sysdrain-offline.sh <vm uuid> <destination hypervisor>
it will migrate a single vm to <destination hypervisor>, the vm will be stopped first.
nova-ip-tenant-per-hv.sh:
USAGE: ./nova-ip-tenant-per-hv.sh <hv ip>
will show which vm run on the hypervisor, their ip address and the security group (which is named after the tenant)
load-script: USAGE copy network information on the second,third, fourth, etc lines and run. it will spawn many as 2 iperf processes peer vm. one client one server. also it will run stress, script is tested on centos.
nova-capacity-available: this script will retrieve hypervisor stats, apply overcommit rules and display how many vm of each size can be boot.
most-common-flavors.py: it list the 10 most used flavors, across a cloud, and how many vm use them.
nova-list-aggregates-for-hv.py list all the aggregates a host is in, passing -D it will show disabled hosts
remaining-ip.py list how many ports in each status a network has and how many ip are allocated on its subnets if you subtract ports to allocated ip you know how many ports can be created. some of the ports mark as DOWN may be recycled but is a manual operation (delete if not in use)
test-new-vlan this script is useful after add a new vlan to neutron, it needs thet network uuid as parameter, it will boot a small vm on each hypervisor (using the network you give) and validate that dhcp and metadata are working, it will print 'OK' or 'FAILED' for each hypervisor.
explanation of the script: nova boot --user-data --> pass user-data to nova
echo "#!/bin/sh" > $TEMP echo echo $name >> $TEMP echo echo $hv >> $TEMP user data, in this case, is a shell script which print the hypervisor and vm name
user data can be only retrieved by the vm over metadata service, cloud-init does that at boot, and executes if user-data is an script
dhcp, metadata and all vm network traffic is "inside" of a vlan, vm does not not it but all the traffic gets tagged by the hv
if the switch port is not configured to allow that vlan tag to pass, neither dhcp, metadata or any other vm traffic will pass
- nova-check-all-computes: will boot a vm on each hypervisor
- nova-images-per-hypervisor: will show all images in use on each hypervisor
- *