summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
| * | | | | | | | | Send a full instance in rollback_live_migration_at_destination.Russell Bryant2012-07-304-17/+24
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Change the rollback_live_migration_at_destination method of the compute rpc API to take a full instance over rpc instead of just the instance ID. This cuts down on database access needed by nova-compute. This was the last method of this API that was taking an instance ID as an argument. There still some left that take the instance UUID, but I can see the light at the end of the tunnel. Part of blueprint no-db-messaging. Change-Id: Id581d23d36fee5c6fc06d2a655e999fb9aae5ae3
* | | | | | | | | | Merge "Convert virtual_interfaces to using instance_uuid."Jenkins2012-07-3116-152/+391
|\ \ \ \ \ \ \ \ \ \
| * | | | | | | | | | Convert virtual_interfaces to using instance_uuid.Michael Still2012-07-3116-152/+391
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This review converts virtual_interfaces from tracking their instances with instance_id to using instance_uuid. This should be the last review for blueprint finish-uuid-conversion. Change-Id: I303bba009fa37c72394369c8338e6e759f794cd0
* | | | | | | | | | | Merge "Turn on base image cleanup by default."Jenkins2012-07-312-2/+2
|\ \ \ \ \ \ \ \ \ \ \
| * | | | | | | | | | | Turn on base image cleanup by default.Michael Still2012-07-282-2/+2
| | |_|_|_|_|/ / / / / | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Essex shipped with the libvirt image cache manager hobbled so it wouldn't delete unused base images. Its been a while and we haven't had a new bug reported in ages. Let's turn it on for Folsom. This is tracked by bug 1029674. Change-Id: I5529232bf4a316f844d383d5f5a3c0c287d59c1f
* | | | | | | | | | | Merge "Compute: VM-Mode should use instance dict."Jenkins2012-07-312-11/+6
|\ \ \ \ \ \ \ \ \ \ \
| * | | | | | | | | | | Compute: VM-Mode should use instance dict.Rick Harris2012-07-302-11/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Instance dicts no longer support attribute-style access. Change-Id: Ia5191aa14ad7cc62630bd6b0df8aaccb87e51e03
* | | | | | | | | | | | Merge changes I55bdc631,I57789eab,I5ed8707b,Iaae16e63Jenkins2012-07-314-71/+96
|\ \ \ \ \ \ \ \ \ \ \ \ | | |_|_|/ / / / / / / / | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * changes: Send a full instance in revert_resize. Send a full instance in resume_instance. Send a full instance in resize_instance. Send a full instance in reset_network.
| * | | | | | | | | | | Send a full instance in revert_resize.Russell Bryant2012-07-304-18/+25
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Change the revert_resize method of the compute rpc API to take a full instance over rpc instead of just the instance UUID. This cuts down on database access needed by nova-compute. Part of blueprint no-db-messaging. Change-Id: I55bdc63149c0385fbf151f4f86f40e6c7635cf63
| * | | | | | | | | | | Send a full instance in resume_instance.Russell Bryant2012-07-304-17/+20
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Change the resume_instance method of the compute rpc API to take a full instance over rpc instead of just the instance UUID. This cuts down on database access needed by nova-compute. Part of blueprint no-db-messaging. Change-Id: I57789eab047de7ff9b84e2bb5367139a9a7b8e40
| * | | | | | | | | | | Send a full instance in resize_instance.Russell Bryant2012-07-304-30/+35
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Change the resize_instance method of the compute rpc API to take a full instance over rpc instead of just the instance UUID. This cuts down on database access needed by nova-compute. Part of blueprint no-db-messaging. Change-Id: I5ed8707b2d018559ab4e983f8528dac8377783c7
| * | | | | | | | | | | Send a full instance in reset_network.Russell Bryant2012-07-304-15/+25
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Change the reset_network method of the compute rpc API to take a full instance over rpc instead of just the instance UUID. This cuts down on database access needed by nova-compute. Part of blueprint no-db-messaging. Change-Id: Iaae16e63da9d48d86e3bfc80f6c0cdb9f9bd3ddd
* | | | | | | | | | | | Merge "Send a full instance in rescue_instance."Jenkins2012-07-314-20/+30
|\| | | | | | | | | | |
| * | | | | | | | | | | Send a full instance in rescue_instance.Russell Bryant2012-07-304-20/+30
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Change the rescue_instance method of the compute rpc API to take a full instance over rpc instead of just the instance UUID. This cuts down on database access needed by nova-compute. Part of blueprint no-db-messaging. Change-Id: I559bd45a8dda4537f6f4a3b8d744095291c7f058
* | | | | | | | | | | | Merge "Send a full instance in remove_volume_connection."Jenkins2012-07-314-14/+22
|\| | | | | | | | | | |
| * | | | | | | | | | | Send a full instance in remove_volume_connection.Russell Bryant2012-07-304-14/+22
| | |_|_|_|_|_|_|/ / / | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Change the remove_volume_connection method of the compute rpc API to take a full instance over rpc instead of just the instance ID. This cuts down on database access needed by nova-compute. Part of blueprint no-db-messaging. Change-Id: I1bd47b161acd7b6dfb03adb77d00a861f754e188
* | | | | | | | | | | Merge "Fix rootwrapper with tgt-admin"Jenkins2012-07-312-4/+4
|\ \ \ \ \ \ \ \ \ \ \ | |/ / / / / / / / / / |/| | | | | | | | | |
| * | | | | | | | | | Fix rootwrapper with tgt-adminChuck Short2012-07-302-4/+4
| | |_|_|_|_|_|_|/ / | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Dont use the full path when calling tgt-admin. LP: #1030850 Change-Id: Ifc7a1960aa67325e9219309e06de292010d7ec46 Signed-off-by: Chuck Short <chuck.short@canonical.com>
* | | | | | | | | | Merge "Fix image_type=base after snapshot"Jenkins2012-07-311-1/+7
|\ \ \ \ \ \ \ \ \ \ | |_|_|_|/ / / / / / |/| | | | | | | | |
| * | | | | | | | | Fix image_type=base after snapshotKevin L. Mitchell2012-07-301-1/+7
| | |/ / / / / / / | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When an instance is snapshotted, the resultant image should have the image_type property set to "snapshot". However, if the base image already had image_type set to "base", the image_type on the snapshot will also be set to "base". This turns out to be because we set image properties in two different places. The first is when the image record is created in glance, which occurs in nova.compute.api:API._create_image(); this is when image_type is set to "snapshot". However, in nova.virt.xenapi.vm_utils:upload_image(), all the properties from the base image are copied over, so all the properties set by _create_image() could be overwritten. The quick solution was to extend the non_inheritable_image_properties configuration setting to also cover the image properties set by _create_image(). Long-term, these two metadata-setting processes need to be more closely unified. (It should be noted that the upload_image() function contains a TODO stating that the code should move into the compute manager, so it can be shared across hypervisors…) Fixes bug 1031079. Change-Id: I936bc147fb823eca67e40084cba5c3bc1a7fe73a
* | | | | | | | | Merge "Send a full instance in remove_fixed_ip_from_instance."Jenkins2012-07-304-16/+22
|\ \ \ \ \ \ \ \ \
| * | | | | | | | | Send a full instance in remove_fixed_ip_from_instance.Russell Bryant2012-07-304-16/+22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Change the remove_fixed_ip_from_instance method of the compute rpc API to take a full instance over rpc instead of just the instance UUID. This cuts down on database access needed by nova-compute. Part of blueprint no-db-messaging. Change-Id: Iee0daf068be9814ef0ba06a69111a45a2e198cbb
* | | | | | | | | | Merge "Send a full instance in rebuild_instance."Jenkins2012-07-306-35/+43
|\| | | | | | | | | | |/ / / / / / / / |/| | | | | | | |
| * | | | | | | | Send a full instance in rebuild_instance.Russell Bryant2012-07-306-35/+43
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Change the rebuild_instance method of the compute rpc API to take a full instance over rpc instead of just the instance UUID. This cuts down on database access needed by nova-compute. This patch also includes fixes for a couple of places that were getting the instance name using the old instance.name syntax. It must be instance['name']. These no-db-messaging patches are making it so more code paths have an instance as a dict (received over rpc) instead of the SQLAlchemy model (looked up from the db on the compute node). Part of blueprint no-db-messaging. Change-Id: I21647a5a5cbf71a2c498f6fee2e4adbe30b8f2ea
* | | | | | | | | Reverts fix lp1031004Matt Dietz2012-07-302-13/+5
|/ / / / / / / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The fix was incorrectly applied to the resize.end event. There was a miscommunication about the resize.end versus resize.finish event Change-Id: I301b46bbb7684d70946a426ce371eaf1198b97df
* | | | | | | | Merge "Resize.end now includes the correct instance_type"Jenkins2012-07-302-5/+13
|\ \ \ \ \ \ \ \
| * | | | | | | | Resize.end now includes the correct instance_typeMatt Dietz2012-07-302-5/+13
| |/ / / / / / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Fixes bug 1031004. Moves the notification to the destination compute node so that the correct instance_type attributes are assigned. Change-Id: I5886154ddd7a9768dc5d82dc4e43e9e0f681ee04
* | | | | | | | Merge "Avoid error during snapshot of ISO booted instance"Jenkins2012-07-302-3/+15
|\ \ \ \ \ \ \ \ | |_|_|_|/ / / / |/| | | | | | |
| * | | | | | | Avoid error during snapshot of ISO booted instanceMark McLoughlin2012-07-302-3/+15
| | |_|/ / / / | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Fixes bug #1023492 If you boot an instance from an image with disk_format=iso and then attempt to snapshot (i.e. nova image-create) you currently get a traceback because we do: qemu-img convert -f qcow2 -O iso -s $snap $disk $out and 'iso' is not a format that qemu-img knows anything about. When booting the image, we use qemu-img to detect that the file is a raw image so we avoid having to special case disk_format=iso there. However, there's no way of avoiding the special casing when extracting a snapshot. Note 1 - it's not very clever to take a snapshot of a read-only disk downloaded from glance and upload it back to glance again. Adding such smarts would be a nice enhancement. Note 2 - only the destination format is important here because we can only be extracting from a qcow2 image since snapshots only work where use_cow_images=True. See also #1030844. Change-Id: I21ff6db8ebb59a83d27f224283fb76f582c38a0e
* | | | | | | Use common parse_isotime in GlanceImageServiceBrian Waldon2012-07-302-18/+8
| |/ / / / / |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Rather than reimplement openstack-common's isotime functionality, just use what's provided for us! Change-Id: Id44eab3c34ae2c682d27d1563c6baaf68b979d2b
* | | | | | Merge "Make compute only auto-confirm its own instances"Jenkins2012-07-306-22/+46
|\ \ \ \ \ \ | |_|_|/ / / |/| | | | |
| * | | | | Make compute only auto-confirm its own instancesChris Behrens2012-07-286-22/+46
| | |/ / / | |/| | | | | | | | | | | | | | | | | | | | | | | Fixes bug 1030226 Change-Id: I0e0cf0c9950f34b1bdab856b2e1c6872b29fdedc
* | | | | Add a link from HACKING to wiki GitCommitMessages pageDaniel P. Berrange2012-07-301-0/+6
| |/ / / |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The HACKING file contains a short example of an effective commit message. For reasons for space, it cannot describe the rationale behind this example. It also does not have space to describe how to split up a patch into a series of commits. Add a link from the HACKING file to the wiki http://wiki.openstack.org/GitCommitMessages where further information can be found Change-Id: I34d44485486b623b11743106f09d5ef631d35888
* | | | Merge "EC2 DescribeImageAttribute by kernel/ramdisk."Jenkins2012-07-302-0/+22
|\ \ \ \
| * | | | EC2 DescribeImageAttribute by kernel/ramdisk.MotoKen2012-07-252-0/+22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Fixes bug 1026898. Supports kernel/ramdisk attributes for EC2 DescribeImageAttribute API to display the ID of the kernel/ramdisk associated with the AMI. And adds test cases to verify this behavior. Change-Id: I3ea91b95812dcec349b4ff6dc889645a57975278
* | | | | Merge "maint: libvirt imagecache: remove redundant interpreter spec"Jenkins2012-07-303-3/+0
|\ \ \ \ \ | |_|_|/ / |/| | | |
| * | | | maint: libvirt imagecache: remove redundant interpreter specPádraig Brady2012-07-293-3/+0
| | |/ / | |/| | | | | | | | | | | | | | | | | | This causes "non-executable-script" warnings in rpmlint at least. Change-Id: Ia62ce7026cfa59cfc7ab457fbb246496c8f58c10
* | | | Merge "Support external gateways in VLAN mode."Jenkins2012-07-291-2/+4
|\ \ \ \
| * | | | Support external gateways in VLAN mode.mathrock2012-07-281-2/+4
| |/ / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add network label when starting dnsmasq. Fixes bug 1012443. To allow for per-network dnsmasq customization we can use the network label from the database as the dnsmasq network tag. A user can then specify a custom dnsmasq config file using the 'dnsmasq_config_file' configuration flag. An example dnsmasq config file showing how to set a custom external gateway or ntp server for particular networks based on their tag/label follows: dhcp-option=tag:'private',option:router,10.0.0.254 dhcp-option=tag:'private',option:ntp-server,10.0.0.253 dhcp-option=tag:'private',option:dns-server,8.8.8.8 dhcp-option=tag:'blue',option:router,10.10.0.254 dhcp-option=tag:'blue',option:ntp-server,10.10.0.253 dhcp-option=tag:'blue',option:dns-server,8.8.4.4 Change-Id: I56cb175a968e33b803dfcb965f90f3a5ca4983ca
* / / / Check for selinux before setting up selinux.Chuck Short2012-07-291-1/+3
|/ / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Check to see if /etc/selinux exists before trying to setup ssh keys for selinux. If it doesnt exist it will result in a traceback when setting up the image since the directory doesnt exist on Ubuntu. Fixes LP: #1030609 Change-Id: I8af1e6232bc5a84cbb878c2663ab30eb4139484c Signed-off-by: Chuck Short <chuck.short@canonical.com>
* | | Merge "Fix state logic for auto-confirm resizes"Jenkins2012-07-282-4/+7
|\ \ \
| * | | Fix state logic for auto-confirm resizesKevin L. Mitchell2012-07-272-4/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In the compute manager, the _poll_unconfirmed_resizes() periodic task mishandled state checks, and would call confirm_resize() in invalid states. This patch fixes the logic so that confirm_resize() is only called if the instance is in vm_state 'resized' and task_state None. The unit test is also fixed for the proper logic. Fixes bug 1030175. Change-Id: I66028b7042d88fb0f1f41df07f42b4b02cf60247
* | | | Merge "Add persistent volumes for tgtd."Jenkins2012-07-284-75/+98
|\ \ \ \ | |/ / / |/| | |
| * | | Add persistent volumes for tgtd.Chuck Short2012-07-274-75/+98
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently if you restart the server running nova-volume or restart tgt, you will loose your iscsi targets that have been created. This is not good. In order for iscsi targets to be persistent across reboots or restarts, one has to have the target's configuration information in /etc/tgt/targets.conf or /etc/tgt/conf.d. So when tgtd is restarted then the iscsi targets will be there as expected. This patch will add a configuration file to $state_path/volumes when the volume is created. The configuration file is identified by the volume uuid. It creates a logicalunit when the volume is created as well. The iscsi target and configuration file will be removed once the volume has been removed as well. In order to use this, you have to include the following in your /etc/tgt/targets.conf include $state_path/volumes/* For upgrades, it will just re-create the volumes already in the volumes table. Fixes LP: #1011159 Change-Id: I38fc096ab881ccb52cb688ae46d9d36b0a7b3a45 Signed-off-by: Chuck Short <chuck.short@canonical.com>
* | | | Merge "Explicitly send primitive instances via rpc."Jenkins2012-07-271-4/+4
|\ \ \ \
| * | | | Explicitly send primitive instances via rpc.Russell Bryant2012-07-271-4/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This fixes a few bugs in recent no-db-messaging commits. There were a few places that I accidentally still used "instance" where I should have used "instance_p", which is the instance converted to primitive types. I noticed the problem looking at smokestack logs. I saw a Qpid traceback that indicated the problem. As an interesting point, systems using Kombu would not have seen this problem. Kombu uses anyjson. Our jsonutils module sets anyjson to go through jsonutils, which will automatically to jsonutils.to_primitive() if needed to serialize an object. Qpid doesn't do json serialization, so it will blow up if it hits a non-primitive type. Yay for testing multiple different configurations to find problems. Part of blueprint no-db-messaging. Change-Id: I3487195110572b7ae9586cfcc618ab7ab28d5017
* | | | | Merge "libvirt driver: set os_type to support xen hvm/pv"Jenkins2012-07-274-7/+90
|\ \ \ \ \
| * | | | | libvirt driver: set os_type to support xen hvm/pvChristoph Thiel2012-07-264-7/+90
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Generate libvirt config for xen hvm and pv guests based on instance vm_mode. The vm_mode can be set via the vm_mode image property. Use hvmloader for vm_mode.HVM guests. libvirt_type xen still defaults to vm_mode.XEN (pv), to not change any behavior. Change-Id: I5f92bedf6588ab2674d784772a5ecdb895fabe40
* | | | | | Merge "Return location header on volume creation"Jenkins2012-07-274-6/+15
|\ \ \ \ \ \ | |_|/ / / / |/| | | | |
| * | | | | Return location header on volume creationEoghan Glynn2012-07-275-7/+16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Partially addresses LP 1026600 Normal RESTful idiom would dictate that the "Location" header is set to reference the newly created resource. We defer lining up the status code returned with generally accepted RESTful convention (currently "200 OK", should be "201 Created" or "202 Accepted") until the next API major version bump. There is already substantial code duplication between the volumes API and the os-volumes API extension - this will be addressed in a later patch. Change-Id: I166846a4cfea8adc8c156b8ae0e0d288681ac08c
generated by cgit (git 2.34.1) at 2026-04-05 00:22:00 +0000