~newell-jensen/maas/2.2-fix-1694759 : changes

~newell-jensen/maas/2.2-fix-1694759 » Changes from revision 6061

From Revision 6061 to 6042

Rev	Summary	Authors	Tags	Date
6061	Backport trunk r6074: Delete newly allocated RSD node if the Backport trunk r6074: Delete newly allocated RSD node if the assembling process fails.	Newell Jensen		6 years ago
6060	[r=blake-rouse][bug=][author=bjornt] Add a config file for f... [r=blake-rouse][bug=][author=bjornt] Add a config file for flake8 instead of hardcoding it in Makefile. This allows you to run flake8 in your editor with the same settings as 'make lint-py'.	Bjorn Tillenius		6 years ago
6059	[r=mpontillo][bug=][author=bjornt] Remove the workaround for... [r=mpontillo][bug=][author=bjornt] Remove the workaround for running DHCP tests in a container.	Bjorn Tillenius		6 years ago
6058	[r=mpontillo][bug=][author=danilo] Use 'lxc profile show' ex... [r=mpontillo][bug=][author=danilo] Use 'lxc profile show' exit code to check for profile instead of grepping 'lxc profile list' output which has changed format.	Danilo Šegan		6 years ago
6057	[r=mpontillo][bug=][author=blake-rouse] Build all the servic... [r=mpontillo][bug=][author=blake-rouse] Build all the services that MAAS depends on from the archive. Fix rackd and regiond logging. The snap no longer builds all the required services from source, it now pulls all service from the archive. This also helped reduce the size of the snap almost 50%, without doing any work to remove doc and manuals. rackd and regiond processes are now started with twistd --logger which adds the improved logging in MAAS that you see in developer mode and from packaging.	Blake Rouse		6 years ago
6056	[r=danilo][bug=1690781][author=newell-jensen] Raises VirshEr... [r=danilo][bug=1690781][author=newell-jensen] Raises VirshError if the `virsh domcapabilities --virttype kvm\|qemu` command returns empty XML with a suggestion on how to potentially fix the situation.	Newell Jensen		6 years ago
6055	[r=andreserl][bug=][author=blake-rouse] Add new 'maas init',... [r=andreserl][bug=][author=blake-rouse] Add new 'maas init', 'maas config', 'maas status' commands for the MAAS snap. This branch adds the new 'maas init', 'maas config', 'maas status' commands for the MAAS snap. The MAAS snap now starts in 'none' mode with nothing configured. The 'maas init' gives an interactive setup mode to configure the snap, 'maas config' allow the user to get the current configuration or change the snap configuration, and the 'maas status' provides the current status of all services running in the snap.	Blake Rouse		6 years ago
6054	[r=mpontillo][bug=1687420][author=bjornt] When resetting the... [r=mpontillo][bug=1687420][author=bjornt] When resetting the network configuration over the API, remove all network interfaces that weren't detected during commissioning.	Bjorn Tillenius	2.2.0	6 years ago
6053	[r=newell-jensen,andreserl][bug=1688060][author=danilo] When [r=newell-jensen,andreserl][bug=1688060][author=danilo] When resyncing/refreshing a pod, ensure machines turned into rack controllers are not recommissioned.	Danilo Šegan		6 years ago
6052	[r=andreserl][bug=][author=danilo] Fix a few typos found aro... [r=andreserl][bug=][author=danilo] Fix a few typos found around comments, docstrings and method names.	Danilo Šegan		6 years ago
6051	[r=blake-rouse][bug=1689838][author=mpontillo] Fix database ... [r=blake-rouse][bug=1689838][author=mpontillo] Fix database fields to pass in field name when raising ValidationError.	Mike Pontillo		6 years ago
6050	[r=andreserl][bug=1689288][author=ricgard] Resolving edit VL... [r=andreserl][bug=1689288][author=ricgard] Resolving edit VLAN's issue on Machine interfaces	Richard McCartney		6 years ago
6049	[r=andreserl][bug=1690144][author=blake-rouse] Handle block ... [r=andreserl][bug=1690144][author=blake-rouse] Handle block devices in a virsh domain that is missing its backing. When a virsh domain has a storage device thats backing file no longer exists or cannot be accessed no capacity is given for that storage device. Because of this virsh will also refuse to start the domain. MAAS does not add this domain because it cannot be used at all.	Blake Rouse		6 years ago
6048	[r=blake-rouse][bug=1689628,1689629,1689630][author=newell-j... [r=blake-rouse][bug=1689628,1689629,1689630][author=newell-jensen] Fixes several RSD pod driver bugs encountered while trying to compose machines with remote storage. * Passes missing argument to get_pod_machine in compose method * * Re-scrapes the logical volumes and targets after a node is composed so that this information is current for get_pod_machine call in the compose method * * Uses getattr to test if the request's block devices have iscsi_target set *	Newell Jensen	2.2.0rc4	6 years ago
6047	[r=blake-rouse][bug=1686169][author=ltrager] Filter out unsu... [r=blake-rouse][bug=1686169][author=ltrager] Filter out unsupported products from BootSourceCache and send MAAS user_agent. In the previous attempt the filter was applied whenever RepoDumper was used. Both the region and the rack use RepoDumper to list what images are available in a stream. The stream the region produces for the rack is similar, but not identical to the one at images.maas.io. These differences cause all images to be filtered out. Validation now only occurs when the region reads images.maas.io. The rack imports all images from the region and does not do any filtering.	Lee Trager		6 years ago
6046	[r=blake-rouse][bug=1686757][author=danilo] Support rack con... [r=blake-rouse][bug=1686757][author=danilo] Support rack controllers using fully-qualified domain names as their hostnames When a rack controller has a hostname set including a domain name (in the dotted form, eg. foo.example.com), registering this rack controller to existing MAAS regiond should keep the FQDN for the controller (if a domain does not exist already, it is added as non-authoritative domain name in MAAS).	Danilo Šegan		6 years ago
6045	[r=blake-rouse][bug=1688661][author=newell-jensen] Move wher... [r=blake-rouse][bug=1688661][author=newell-jensen] Move where remote storage calculations are done for master_path, master_size, and master_id as this is not dependent on whether or not the logical volume has targets or not. Fix message that is displayed when failing to compose or decompose a machine.	Newell Jensen		6 years ago
6044	[r=ltrager][bug=1687076][author=newell-jensen] Create signal... [r=ltrager][bug=1687076][author=newell-jensen] Create signal to set node.owner to None whenever a machine (not controller) transitions to the NEW or READY state.	Newell Jensen		6 years ago
6043	[r=newell-jensen][bug=1688009][author=ltrager] Ignore timeou... [r=newell-jensen][bug=1688009][author=ltrager] Ignore timeout when telling the rack to disable and stop itself on deletion. When the system running a rack controller is shutdown and a user tries to delete it the region may still think the rack is connected. The region will attempt to send an RPC call instructing the rack to disable and stop itself. This is to ensure the rack doesn't reconnect causing MAAS to recreate a node object for it. If the system is turned off it won't be coming back. Wait for the timeout to occur but shorten it as we really do want the rackd service to be off after deletion.	Lee Trager		6 years ago
6042	[r=newell-jensen][bug=1666997][author=ltrager] Notify and wa... [r=newell-jensen][bug=1666997][author=ltrager] Notify and warn the user when the ephemeral operating system is unavailable. The settings page allows a user to select which operating system MAAS should use during commissioning and other ephemeral environments. This list was built from operating systems that are currently available on rack controllers. If the currently selected operating system is not available the first choice was selected. This led the user to believe a usable operating system was selected when it really wasn't. When the ephemeral operating system is not available MAAS creates a notification for all users, shows an error on the images page, lists the currently selected commissioning operating system as unavailable on the settings page, and an exception is raised when the user tries to use the ephemeral operating. Users are no longer able to remove the current ephemeral operating system from the BootSourceSelection table.	Lee Trager		6 years ago

Older »