Tuskar-UI currently uses the default stack timeout in heat which is
60 minutes. The CLI defaults the stack timeout to 240 minutes.
This patch adds the timeout_mins paramater set to 240 to match the
same behavior as the CLI.
Change-Id: I983db9218e43020235c54856ca6b435ebc33400b
python-ironic-inspector-client works with both ironic-discoverd
and newly introduced ironic-inspector. It also pulls in less
dependencies (aka does not pull in the whole service).
Change-Id: Iaab480b4a2e4556dd6b96b42e4cecdbef878ff45
Previously the dropdown choice selection of possible network
interfaces was limiting. We are now using a free text field
to remove this limitation.
Change-Id: I1eb1e495282e7fa520c8225ba16fa61a32b3d24d
Closes-bug: bz/1252553
If node introspection fails, there is no way to re-run the process
within the UI. This patch adds a button to the maintenance tab
to allow a user to re-run the process.
Change-Id: I16645643de83f212e89f858a85589d425474c5cd
No overcloudrc was created when deployed via the UI. This adds the
ability to view the values and to download an appropriate rc file.
Change-Id: I2ee7cc4b10559a3542a6d22b01179b764c857d6b
This is achieved by removing context_selection templatetag module,
that we used to override horizon behavior
Due to this bug https://bugs.launchpad.net/horizon/+bug/1023502
it is not possible to determine in which tenant the user will get
logged in. So we need to let user change tenant if needed.
Change-Id: Ib7f7076678e6ec51a88dca7bbd9a472adbadaf2a
After deploying an overcloud via the UI, the generic error page for
Horizon is displayed. With debugging enabled, an EndpointNotFound
error message is displayed. This is caused by the dashboard_urls
method attempting to access a method on the overcloud keystone client
before the overcloud keystone service is initialized.
This patch wraps the code to check for initialized services in a
try block and returns an empty list in the event an exception is
thrown.
Change-Id: I7da53ea085b2728950c3a0c2760d0745119e6ff1
Resolves: rhbz#1245192
Wholesale syncing of the parameters across roles doesn't work. There are
some parameters that can be different depending on the role. For one,
NeutronPublicInterface might be different depending on the node's role.
This patch comments out the sync. Future might involve creating a
white/blacklist for which params should/should not be synced.
Change-Id: I66be8126e0160aaa0af2f49b7d5fce705e720f52
Some parameters were missing when controllers > 1 and when Ceph nodes
are in play. This adds those missing params when the plan is edited.
Change-Id: Ieefd9b901f6332be5fff70a349b06d6a2e0d2e2b
The post-deployment init always creates a default-net tenant network,
a router and an ext-net external network. While this is good for demo
purposes they are not needed in a production environment for naming
or functional reasons and first thing that an operator does is to delete them.
You might not need the default router. For instance if you want to test vlan
provider networks. But the init creates the ext-net network and default router
though they are not needed.
It generates confusion when using network isolation. You have to specify
floating ip cidr, floating ip start, floating ip end, bm network gateway,
that match the external network settings set up in the network isolation
yaml file in order to get working connectivity. And you may end up deleting
them right after init finishes so you get a bad user experience.
Closes-bug: https://bugzilla.redhat.com/show_bug.cgi?id=1243056
Change-Id: Iddfcc39498ce1ac991b103df0b32b49f074deb34
Through additional profiling, I found a few more places where
caching could be enabled to reduce network calls and speed up
page load times slightly. This patch adds caching to frequently
used methods and provides for minor improvements. Tested in both
virt and baremetal environments.
Change-Id: I768d6643c4523d33c440a13f5ddc5212555956a8
Handling of provisioning images has changed on the Tuskar side;
this fix brings Tukskar-UI in line with these changes.
* There used to be separate images for all the deployment roles.
This is no longer the case and all the roles now use the same
image, 'overcloud-full'. As the result, the way the relationship
between images and roles is displayed on the Provisioning Images
page needs to be updated.
* Related to the previous point, test data needed updating to
reflect that.
* Images associated with roles are now saved by name in the Tuskar
plan. This fix ensures that the UI saves and retrieves images
by name and not by UUID, like it used to do.
Change-Id: I11b73fcaed45200dfbdb72dc3ab9a27afb67a31d
Page load times in tuskar-ui are too slow. This patch adds some
caching to the tuskarclient and its list function to reduce page
load time. Tests locally showed these changes decreased page load
time by half in a virt environment.
Change-Id: Ib5d67dc2c1a58011739b81156fa50f5cccb223d4
We used to filter by image_id for roles, back when we had different
images for different roles. As we are now using the same image for
all the roles, that way of filtering does not work any longer. We
need to filter by instance UUIDs instead.
Change-Id: Ie4f26a29beaaf67aa3b36bb82961b491d31c95a0
The Role resource group name now matches role name, so there is no
need to specify provider_resource_group_name property on Role
Change-Id: Ibccc3b65e63450feec894be7abc6551fc75b4ed4