CloudStack uses Guest CIDR for dhcp-range for the Guest VMs. The entire
CIDR is used by CloudStack for assigning IPs to Guest VMs. IP Address
Reservation will allow part of address space to be used fornon CloudStack
hosts/physical servers also, by restricting the address space of CloudStack
Guest VMs. Reservation can be configured using update Network API by specifying
guestvmCidr as an additional parameter. Reservation will be applicable for
Isolated Guest Networks including VPC. reservediprange in the response
will return the IP range that can be used for non Cloudstack hosts.
Tested manually the following scenarios:
Applying reservation when there are running VMs inside the
guest_vm_cidr.
Applying reservation when there are running VMs outside the
guest_vm_cidr.(not allowed)
Applying reservation when external device like Netscaler is configured
in the guest_cidr.
Applying reservation in VPC tiers.
Applying reservation outside the range of guest_cidr.(not allowed)
Supporting kickstart in CloudStack baremetal
able to start vm
Conflicts:
client/tomcatconf/componentContext.xml.in
server/src/com/cloud/baremetal/BareMetalTemplateAdapter.java
server/src/com/cloud/baremetal/BareMetalVmManagerImpl.java
server/src/com/cloud/vm/UserVmManagerImpl.java
There was previously a limitation in the detail view/quickview where
only 1 'view all' link was visible per page. Change this to show n
number of view all links, if an array is passed to 'viewAll.' This also
lowers the # of clicks necessary to navigate to some items.
This also adds more view all buttons to instances, storage, and system
resources.
It's a regression caused by:
commit 69d24545c4f0e316dba534b6d5d99308795db6d4
Author: Murali Reddy <murali.reddy@citrix.com>
Date: Fri Feb 8 18:50:18 2013 +0530
CLOUDSTACK-1208: Failed to shutdown guest network
Firewall manager was being used instead of LoadBalancingRules manager
while applying the load balancer rules in shut down network. Changing it
to LoadBalancingRules manager.
The parameter transferre to applyRules() should be lbRules, rather than the lbs
object.
new call to network.getPhysicalNetworkId() which wasn't mocked yet in
the nvp plugin unittests.(cherry picked from commit aea5b268b4590775eff6291ddfbbd6de777d1b63)
Signed-off-by: Hugo Trippaers <trippie@gmail.com>
In c63dbb88042d7eabea2664c2b608c51792fc9f18 I removed the rule from create-schema:
- `queue_proc_time` datetime COMMENT 'when processing started for the item',
But, upgrade path schema-40to410.sql had a different rule which caused the bug:
+ALTER TABLE `cloud`.`sync_queue_item` ADD `queue_proc_time` DATETIME NOT NULL
COMMENT 'when processing started for the item' AFTER `queue_proc_number`;
In this fix we just revert to whatever rule was defined in create-schema as the
developer may have forgetten to fix same rule in create-schema and upgrade path.
This commit can be reverted or the code be fixed if we want that queue_proc_time
cannot be null.
Signed-off-by: Rohit Yadav <bhaisaab@apache.org>
