A possibility exists where inspector *can* fail upon inspection if
the database connectivity was lost on a prior action. This is because
the last database transport is potentially bad and fails upon load for
the transaction. The cache can then end up with an "error" state entry,
which upon retrying can fail becasue it is already in error state

Because there really are no guarentees regarding database failures,
the best thing to do is to not trust the prior cache state if it is
in error and to reset it to starting upon new introspection requests.
This prevents operators from *having* to perform process restarts to
force all loads to be from the database unless they manage to have a
multi-inspector cluster and get another inspector node to inspect in
the mean time.

Change-Id: I04ae1d54028862642d043f3a8f3af99405863325
Story: 2008344
Task: 41246
Related: rhbz#1947147
(cherry picked from commit d972dc93)

As discussed and approved during the ironic wallaby midcycle, the grenade
jobs in old stable branches, starting from EM branches and up the oldest
not-EM (currently ussuri) are removed.

Change-Id: I61fc181cb488e913f828efa18dc010348986380d

We are using sphinx latest verion (3.5.0), this causes
`Extension error (sphinx.ext.viewcode)` [1].
The problem was found in [2]
Adding ussuri requirements to docs env will avoid the problem.

[1] https://github.com/sphinx-doc/sphinx/issues/8880
[2] https://review.opendev.org/c/openstack/ironic-inspector/+/775640

Change-Id: I1f03b4ba42025c185f530a9a2a564056c4fa405f

Since pymemcache 3.4.0, localhost will first translate to ipv6 ([::1])
if possible instead of ipv4 (127.0.0.1). Since memcached is running in
the ipv4 interface, we need to point tooz to the right endpoint.

Conflicts:
  zuul.d/ironic-inspector-jobs.yaml

NOTE(moguimar): Conflicts are due small difference in adjacent lines.

Signed-off-by: Moisés Guimarães de Medeiros <moguimar@redhat.com>
Change-Id: Iddf558cd598e41ffead11e164ce2fb1720d15f34
(cherry picked from commit 86b974b7)
(cherry picked from commit 965a4f50)

The pattern of adding a column and then reading a table with it
no longer works in SQLAlchemy 1.3.20. This has been reported
upstream [1] and apparently we're doing the wrong thing by reusing
a column object twice.

We need to disable the non-standalone job since it's really broken
now, and this fix is blocking bifrost.

Also remove lower-constraints job.

Plus increase memory for grenade job.

And since tinyipa is not that tiny anymore, we increase the memory
for the base job to 512.

[1] https://github.com/sqlalchemy/sqlalchemy/issues/5669

Change-Id: I2fb07413e8f421f39b24acf1272771ee2097b195
(cherry picked from commit 5678f219)
(cherry picked from commit c6fdf251)
(cherry picked from commit 1884ba81fd3d6166c626afd008beddaf0a33801e)

Fixes the ironic utilities logic to parse URLs in driver_info
properly.

Also adds ``redfish_address`` to the default list of fields to
evaluate for host matching.

Previously an operator could add the field to the list, but the
matching logic would not match the URL as it did not know to
decompose the url and identify the hostname portion of the url.

Conflicts:
      ironic_inspector/conf/default.py

Resolves: rhbz#1670336
Story: 2008010
Task: 40660
Change-Id: Ice86e9ab3efb98b649141bdf7e1e2febdc9203a8
(cherry picked from commit bbc5cd7d)

Inspection can sometimes take longer based on multiple factors.
This patch adds a variable to control the node inspection timeout.

Change-Id: Iefac618d73ccba3b58baefdc6b993758622b8811
(cherry picked from commit a02fcd1e)

Apparently, virtual floppies on some machines are represented as normal
block devices, but with size = 0. We should filter them out in ironic-lib,
but for now at least don't fail miserably.

Story: #2007907
Task: #40308
Change-Id: Ia96a9a8b85841612fb13616754bfdbf651b212d8
(cherry picked from commit 95e103d2)

Uncompressing the ramdisk on a single core CPU can take a
substantial amount of time and elongate the overall initial
period between VM start and OS becoming operational.

Additionally, multiple CPUs allows for the software with-in
the job to execute more quickly with multiple running processes.

Change-Id: I814b12cbc05c6a110dc5593113b948e3f89e80c1

This function does not have a compatibility shim that adds 'uuid'.

Change-Id: I131c8189b25ee2de66f08c742a21052ef1b266aa
(cherry picked from commit 06390e37)

Link local address is a valid address on IPv6 network when we are using
SLAAC instead of DHCPv6, current interface validation bypass link local
address and fails the introspection.
Interfaces without link carrier don't have link local addresses so this
should be fine.

Change-Id: I4cf195fc8906a2d91989df3f588e1e96ca9684b1
Story: 2007816
Task: 40094
(cherry picked from commit 71666178)

Change-Id: I8e46835c56995321c49af5e608b0603c25571148
(cherry picked from commit 8a02e5a5)

Most of the failures in this job is because tempest couldn't
finish the test on a given time, we increased the introspection
timeout but we forgot to allow tempest to run for more than
20 min (default)

Change-Id: I25435fe93f2f6d6cd50befe9cdf6bb1e548cef88
(cherry picked from commit 06ee5007)

Update the URL to the upper-constraints file to point to the redirect
rule on releases.openstack.org so that anyone working on this branch
will switch to the correct upper-constraints list automatically when
the requirements repository branches.

Until the requirements repository has as stable/ussuri branch, tests will
continue to use the upper-constraints list on master.

Change-Id: I298ede3ba207cb460b4708ee095b12bb19673d1d

We've been historically using endpoints without trailing slashes in
our API. Apparently, some libraries (like gophercloud) are quite
opinionated about it (see the story), so let's handle both.

The implementation could be simpler if we just added trailing slash
to all routes, but it would cause redirects for current users.

Change-Id: Icbd971a8e792f93f9c3fa66ba29bec055dcdee32
Story: #2007660
Task: #39749

Without manage_boot the only real action we do is updating the firewall,
which is supposed to be pretty fast. Making the call synchronous means
that ironic will receive information about PXE filter failures and will
not attempt to power on a failed node or before the PXE filter is updated.

Change-Id: I95b63d4e1d4c2c186c0e8234c4382b00aa918aa8
Story: #2007658
Task: #39746
(cherry picked from commit f3e27c19)

When boot is managed by the ironic side, a node is powered on
right after starting introspection. Inspector's own delays play
no role and may actually prevent introspectin from happening if
a node boots faster than it's whitelisted in the PXE filter.

This changes moves the delay handling later in the process and
only does it when manage_boot is True.

Conflicts:
	ironic_inspector/test/unit/test_introspect.py

Change-Id: If7de8b66ea42eff2966c62a9a0529ab9a5c06f26
Story: #2007658
Task: #39745
(cherry picked from commit 3d1bf55b)

Monkey patch the original current_thread to use the up-to-date _active
global variable. This solution is based on that documented at:
https://github.com/eventlet/eventlet/issues/592

Change-Id: If5c4fc8dd8f3dfe87baf1d10a21ff4333cb00e42
Story: 2007614
(cherry picked from commit 1b972c75)

This commit takes example local.conf from docs.openstack.org/ironic
and adds settings needed to run ironic-inpector.

Change-Id: Id1f4f1140526e27a1072181973a9ff695560c97b
(cherry picked from commit 8211ab46)

- Add ipa-builder to grenade
- Disabled smoke tests since they are not lightweight when
using Ironic.
- Using test_baremetal_introspection for the tempest run
pos-upgrade.
- added LIBVIRT_STORAGE_POOL_PATH since some cloud providers
may not have enough disk space.
- lowered the number of VMs to 3 (we don't have many tests)

Change-Id: Ifee10541f369d5bc850ce3d92305810cd3eedd34
(cherry picked from commit bb71167c)

Change-Id: I5b2cede04ec2b76cbdd5ede192fa0eef833afefa
(cherry picked from commit fe15db0e)

It's nice to know when a periodic task fails, but aborting the whole
process makes it fragile in the events of e.g. problems with accessing
ironic, even if transient.

Simplify the overly complex unit tests and split the dual-purpose task.

Story: #2007622
Task: #39653
Change-Id: I64f5ca6e377a4cdfa08319ef9973b591cc552121
(cherry picked from commit 6e38efff)

I don't even remember why we put them there in the first place..

Change-Id: Id502007917ad7dc64a541a455a1092f2334fdbad
(cherry picked from commit d580714c)

Change-Id: I79eee9e04d1c9867e01f0b248432bb13b6f08904

Based on the native 'grenade' job

Depends-On: https://review.opendev.org/548936
Change-Id: I606f8ffed0defe4ebc7fc2ceda1f6853c36a1b05

After upgrade it seems we don't have enough REMAINING_TIME
to finish the smoke run, let's try increassing the
DEVSTACK_GATE_TIMEOUT.

Change-Id: I61b3a1376cfe3575e01e3eb0a877e76456ba5c61

This is a follow up to comments on https://review.opendev.org/720010.
Adding autospect=True requires an update to ironic_inspector.test.\
base.BaseTest.assertCalledWithPatch.

Also add autospec=True in several mocks where it was not used.

Change-Id: Ia340d7a8ecde1c8cf191555de87b11b19b30f06c

- Removes confusing comment in _get_physnet_patch method.
- Move binary operator to follow code style.
- Removed info log entry in before_update.
- Move releasenote to 'other' section as this is no user
  faceing change.

Change-Id: I53d8bf97a87ae365a085a795901eaed1f0c32d9b