Removed conf files from /etc/pmon.d/
as they are being moved to another location.
This is part of an effort to allow pmon conf files
to be selected at runtime by kickstarts.
The change is debian-only, since centos support
will be dropped soon.
Centos' pmon conf files remain in /etc/pmon.d/
Test Plan:
PASS - deb doesn't install anything to /etc/pmon.d/
PASS - rpm files unchanged
PASS - AIOSX unlocked-enabled-available
PASS - Standard 2+2 unlocked-enabled-available
Story: 2010211
Task: 46306
Depends-On: https://review.opendev.org/c/starlingx/metal/+/855095
Signed-off-by: Leonardo Fagundes Luz Serrano <Leonardo.FagundesLuzSerrano@windriver.com>
Change-Id: I086db0750df5626d2a8ba1010153ce4f45535ca5
The current kickstart is not adding starlingx to the ostree remote
branches for worker and storage nodes (non-controller) nodes.
This update catches this case and adds the starlingx branch
to the ostree remote setup.
With this update: /ostree/repo/config
[remote "debian"]
url=http://pxecontroller:8080/feed/rel-??.??/ostree_repo
branches=starlingx;
Test Plan:
PASS: Verify standard system install
PASS: Verify All-in-one system install
PASS: Verify change on both worker and storage node installs
Story: 2009968
Task: 46260
Change-Id: Idc06877e15c412b4f561a9364cf7f0874ab9f9fc
There is a typo on the path
Story: 2010211
Task: 46259
Signed-off-by: Andre Fernando Zanella Kantek <AndreFernandoZanella.Kantek@windriver.com>
Change-Id: I78775801f7e288c1452cf9da54c50bf918649655
In cases when a previous installation has a different partition layout
and the persistent backup partition does not align, make sure we
recreate the persistent backup partition.
This now covers scenarios when:
- The persistent backup partition is discovered but on a different
partition than expected
- The filesystem on the aligned and expected persistent backup
partition is missing or not compliant with ext4
Test Plan:
PASS - Successful install with misaligned partition schemes: 22.06
CentOS -> 22.12 Debian
PASS - Successful install with an aligned partition scheme but missing
filesystem
Change-Id: I44b9c24ae6d220eefa617304b49d7f984b73cf22
Story: 2009303
Task: 46190
Signed-off-by: Robert Church <robert.church@windriver.com>
This correction:
1) removes the localhost addresses from the loopback
config file. Localhost addresses are automatically set by linux
in the loopback interface.
2) corrected the post-up operation for "$mgmt_dev" and
"vlan$mgmt_vlan" devices
3) set the correct client identifier for DHCP
Test Plan
[PASS] Install AIO-DX on IPv6 network, it is correctly
setting the mgmt ip address after first boot on controller-1
Story: 2009968
Task: 46182
Signed-off-by: Andre Fernando Zanella Kantek <AndreFernandoZanella.Kantek@windriver.com>
Change-Id: I0e50e0c85c69ff3a4b7dc1ed4d13e17f750d70d0
Support persistent_size boot option used in install
values during 'dcmanager subcloud add' operation.
- Handle persistent_size increase from previous.
- Disallow persistent_size decrease from previous.
Other changes:
- Drop into shell on boot failure (avoid kernel panic).
This change is pulled in from kickstart.cfg
- Add extra wait for interface to settle after adding
route. During testing it is found that we need to
wait longer here before proceeding.
Test Plan
PASS:
- Boot with existing partitions
- Boot with non-existing partitions
- Boot with default persistent_size setting
- Boot with persistent_size larger than previous
- Partition size is extended
- Filesystem is resized
- Boot with persistent_size smaller than previous
- Warning message is printed
- persistent_size setting is ignored
- Previous partition and filesystem sizes are used
Story: 2010118
Task: 46144
Signed-off-by: Kyle MacLeod <kyle.macleod@windriver.com>
Change-Id: I9fc9a49a15e36e5dd5d56907f1019cf219882ae5
The install value no_check_certificate tranlates into
boot argument inst.noverifyssl=True from subcloud_install.py
This commit ensures that when we see this boot argument
we apply the tls-permissive config value to the repo remote.
Test Plan:
PASS:
- Install on system controller with https_enabled=True
- Without no_check_certificate=True:
- Install fails with 'Unacceptable TLS certificate'
error during ostree pull
- With no_check_certificate=True:
- Install is successful
- ostree pull is using https url
- Install on system controller with https_enabled=False
- ostree pull is using http url (no SSL)
- Install is successful
Story: 2010118
Task: 46180
Change-Id: I162904e52cc63c1a9e4e9f68c6c9921c8c4d8f3a
Signed-off-by: Kyle MacLeod <kyle.macleod@windriver.com>
Due to the changes
bd9e560d4b
which removed the sm-watchdog, we also need to remove residues in
kickstart config.
Story: 2010087
Task: 46007
Signed-off-by: Davi Frossard <dbarrosf@windriver.com>
Change-Id: I17911773ec4db1549df32a77acd43cd4615b28ee
Created a duplicate install of /etc/pmon.d/*.conf files
to /usr/share/starlingx/pmon.d/
This is part of an effort to allow pmon conf files
to be selected at runtime by kickstarter.
Test Plan:
PASS: duplicate conf on deb
Story: 2010211
Task: 46112
Signed-off-by: Leonardo Fagundes Luz Serrano <Leonardo.FagundesLuzSerrano@windriver.com>
Change-Id: Ie07c1bfa370da5b2ec71fe3fce948d59be1dd098
This is part of the change to replace nslcd with sssd to
support multiple secure ldap backends.
This change added pmon configuration file for sssd so that it
is monitored by pmon.
Test Plan on Debian (SX and DX):
PASS: Package build, image build.
PASS: System deployment.
PASS: After controller is unlocked, sssd is running.
PASS: ldap user creation by ldapadduser and ldapusersetup.
PASS: ldap user login on console.
PASS: ldap user remote login by oam IP address:
ssh <ldapuser>@<controller-oam-ip-address>
PASS: ldap user login by local ldap domain within controllers:
ssh <ldapuser>@controller
PASS: For DX system, same ldap functions still work properly after
swact.
PASS: Kill sssd process, verify that it is brought up by pmon.
Story: 2009834
Task: 46064
Signed-off-by: Andy Ning <andy.ning@windriver.com>
Change-Id: I701a4cbbda0f900dafd0456aad63132b62d8424f
This work is part of Debian integration effot.
This work is part of allowing Standard type deployment to unlock
worker nodes.
After unlocking the worker node we observe pmon requesting a reboot
because some services are failed. Services come from packages that are
filtered out on worker node on CentOS.
On CentOS we installed personality based packages on specific
personality type nodes.
On Debian selecting what is needed for personality based packages is
done at runtime.
First packaging needs to be updated for the few pacakges that provide
pmon configuration files, then kickstarts updated to handle these at
runtime.
Add this temporary configuration until packaging is updated.
Tests on Debian:
PASS: controllers unlocked on Standard
PASS: worker node unlocked
no restart initiated by pmon
Story: 2010211
Task: 46025
Signed-off-by: Dan Voiculeasa <dan.voiculeasa@windriver.com>
Change-Id: I01dd7525944a1de919f642df5eceb0b5386cc9af
The initial ostree pull is directed into a mounted
/sysroot directory. During the nochroot stage the
pulled /sysroot repo is subsequently moved
into the rootfs and the ostree remote is reconfigured.
Related to: https://review.opendev.org/c/starlingx/metal/+/852032
Test Plan:
PASS:
- verify subcloud install and bootstrap via
dcmanager subcloud add
- verify single network pull for ostree
Story: 2010118
Task: 46037
Signed-off-by: Kyle MacLeod <kyle.macleod@windriver.com>
Change-Id: Ie250774c8960f7f7ff4a3a62fa98e21547132373
After performing an analysis of the system it was recognized
that the following package: python-ryu; is not being used anymore
by the system. In order to clean it up, it was decided to
remove the package.
Test Plan:
PASS - All pkgs built successfully after removal of python-ryu
PASS - Successfully generated an openstack tarball
PASS - Stx-Openstack tarball successfully applied
PASS - Built CentOS ISO with the change and applied it to a lab
Partial-bug: #1985091
Depends-On: https://review.opendev.org/c/starlingx/integ/+/852863
Signed-off-by: Rafael Cardoso Pereira <rafael.cardosopereira@windriver.com>
Change-Id: I27d119c11d18c1ebbd0beafe154a32960c101529
This commit enables the creation of rvmc docker image by both
distros: CentOS and Debian.
Test Plan:
PASS - Install DC subcloud using Debian image
1. Build rvmc image using stx-debian.
2. Push the new rvmc image to a DC lab registry.
3. Modify the dccommon subcloud_install pointing to the new rvmc image.
4. Install a new subcloud using the new rvmc image.
PASS - Install DC subcloud using CentOS image (same steps as Debian)
Story: 2009831
Task: 45969
Signed-off-by: Hugo Brito <hugo.brito@windriver.com>
Change-Id: I11a9f331cfd1a03d7b5a99223203b240fb31d88b
1. Copying ostree_repo to the feed
The ostree_repo is required at /var/www/pages/feed/<rel-id>
Also, support for IPv6 must be added.
The required support for this is added to miniboot.cfg.
The ostree repo is pulled from the remote to the feed directory
in archive mode.
2. Setup networking for both the ostree_repo pull on initial boot and
the persistent configuration for the subsequent phase 2 boot.
Note: This commit also incorporates and supercedes the changes from
https://review.opendev.org/c/starlingx/metal/+/850237
Test Plan:
PASS:
- ostree is able to install and backup the repository with the provided
network settings.
- Full boot and dcmanager subcloud add via sushy emulator in libvirt
environment (IPv4)
- subcloud is added with appropriate install values and bootstrapped
via the generated bootimage.iso
- Full boot and dcmanager subcloud add via redfish in hardware lab
using IPv6 (Dell hardware only)
Story: 2010118
Task: 45924
Depends-On: https://review.opendev.org/c/starlingx/distcloud/+/851679
Signed-off-by: Kyle MacLeod <kyle.macleod@windriver.com>
Change-Id: Icd1a50985e29841a784c9d339b8717041735c9cf
Sensor and sensorgroup suppression was temporarily disabled in
Debian while System Inventory was modified to align API types
with database types.
That update is now merged so this update removes the Debian only
gate on sensor and sensorgroup suppression.
Test Plan:
PASS: Verify Debian build and install
PASS: Verify CentOS build and install
PASS: Verify multiple individual sensor suppression/unsuppression
PASS: Verify sensorgroup suppression/unsuppression
PASS: Verify host degrade and alarm mgmt for each above 2 cases
Story: 2009968
Task: 45964
Signed-off-by: Eric MacDonald <eric.macdonald@windriver.com>
Change-Id: I34abf6bd7c72df2f7da743e4f20300956248c6d7
This update implements kickstart changes required to install
controller nodes with a single network ostree_repo pull.
Test Plan: Verify single network pull for all
PASS: Verify AIO provisioning from a USB install
PASS: Verify AIO provisioning from a PXE install
PASS: Verify Standard 2+1 System provisioning from a USB install
PASS: Verify worker only install
PASS: Verify controller-1 install from controller-0 (5+)
PASS: Verify Standard controller-0 install from controller-1
PASS: Verify Standard worker install from controller-1
Story: 2009968
Task: 45932
Signed-off-by: Eric MacDonald <eric.macdonald@windriver.com>
Change-Id: I72684d65237c3596ef07a8fdcfc0555fc9e95a3e
This work:
- is part of Debian integration effort.
- will be part of a series of commits.
- will allow puppet manifests to be applied on worker node at
unlock/reboot time.
This specific commit:
- focuses on Standard deployment type
- allows controllerconfig and workerconfig-standalone packages
to coexist on the same iso
- masks controllerconfig and/or workerconfig services based on
personality at runtime
- selects and links hardcoded paths to personality based paths
Future work will account for storageconfig package and service.
Future work will account for other packages.
PASS: controllers unlocked on Standard
PASS: workerconfig started puppet manifest apply on worker node
PASS: controllers unlocked on AIO-DX
PASS: controllerconfig is running on AIO-DX,
controllerconfig is running on Standard,
workerconfig is not running on controller node AIO-DX & Standard,
workerconfig is running on worker node
Story: 2010211
Task: 45949
Signed-off-by: Dan Voiculeasa <dan.voiculeasa@windriver.com>
Change-Id: I548cb4e42be3b9280040d235d67639b5ed4e0de2
This work is part of Debian integration effort.
This work only affects Debian.
This work will work around an issue seen during puppet apply on worker
nodes.
This obsoletes workaround [1]. In fact we move the logic here such that
we don't require manual intervention anymore.
Tests on Debian:
PASS: both controllers unlocked on a Standard deployment.
PASS: puppet manifest applied on worker node.
PASS: both controllers unlocked on an AIO-DX deployment.
PASS: swact on AIO-DX by reboot
[1]: https://review.opendev.org/c/starlingx/utilities/+/850541
Related-Bug: https://bugs.launchpad.net/starlingx/+bug/1983580
Signed-off-by: Dan Voiculeasa <dan.voiculeasa@windriver.com>
Change-Id: Idf49b9e19cc733672b7b878b25563ac6ef9d1d67
The kickstart is not quering system inventory for the
management vlan if the boot interface name lookup succeeds.
This update modifies the logic to perform the system inventory
vlan query for system node installs even if the boot interface
name lookup succeeds.
LAT does not support pxecontroller address translation in a
chroot environment and post chroot hooks are executed before
post nochroot hooks. Therefore, this update had to move the
entire interface setup kickstart segment to a nochroot hook.
Note: This change does not impact CentOS env.
Test Plan:
PASS: Verify Debian build and Install in vbox and RealHw
PASS: Verify platform.conf is updated with vlan<id> if system
inventory mgmtvlan query returns a vlan id.
PASS: Verify interface.d interface setup files are correct
for ipv4 with and without vlan.
Regression:
PASS: Verify system host-reinstall controller-1
PASS: Verify system host-swact controller-0
PASS: Verify system host-lock controller-0
PASS: Verify system host-reinstall controller-0
PASS: Verify system host-unlock controller-0
PASS: Verify platform.conf is updated with management interface
name if system inventory mgmtvlan query returns nothing.
Depends-On: https://review.opendev.org/c/starlingx/tools/+/852029
Story: 2009968
Task: 45895
Signed-off-by: Eric MacDonald <eric.macdonald@windriver.com>
Change-Id: I24eb5d2b2bcc7d9852fda75a27ef9f07e3b75df4
Add upgrades meta files to /upgrades directory in Debian ISO.
Files include:
version
Version file to indicate the ISO release version
metadata.xml
Supportted upgrade path, and required patch
import.sh
Shell script to import the ISO load into StarlingX system
pxeboot-update-<rel-ver>.sh
Shell script to update pxeboot config file for new release
Test cases:
Build Debian ISO and visually inspect the existance of files
list above in /upgrades directory of Debian ISO.
Build Centos ISO with no changes.
Story: 2009303
Task: 45830
Change-Id: I1f0b42c6c37943b19bdf5dd7baaf81da5d1eef7a
Signed-off-by: Bin Qian <bin.qian@windriver.com>
This update adds kickstart support for the following install types
- worker
- worker lowlatency
This update also
- Adds install type prefix to all kickstart logs.
- Fixes an issue that was causing all the kickstart logs produced
with 'ilog' to have the same date.
- Fixes the report_failure_with_msg handling.
- Removed defaulting to aio if traits boot argument is missing.
- Hardens handling of invalid traits and invalid trait combinations.
Install will fail if traits arg is missing or if an invalid
trait combination is specified.
Test Plan:
PASS: Verify Debian Build packages and image
PASS: Verify worker node install and disk boot
PASS: Verify lowlatency worker node install and disk boot
PASS: Verify install type prefix added to logging
PASS: Verify kernel arguements
PASS: Verify no platform backup partition is created on
worker only node
PASS: Verify invalid traits handling
Regression:
PASS: Verify AIO standard install and enable (c0 & c1)
PASS: Verify AIO lowlatency install and enable (c0 & c1)
PASS: Verify standard controller install and enable
PASS: Verify platform-backup partition create and persistence
PASS: Verify system host lock/unlock controller-1
PASS; Verify system host-reinstall controller-1
PASS: Verify logging
Story: 2009968
Task: 45831
Signed-off-by: Eric MacDonald <eric.macdonald@windriver.com>
Change-Id: I6446e5b94b2b386aa0ac1d92e5ffb2cb1167db27
Remove the 20 second delay before bringing up the network
interface. This was done so that the network interface has
feature parity with Centos 7, but ifupdown detects nelink
differently compared to Centos 7. Removing the delay
results in a significant speed up.
In Debian, the network cable test is run by
/etc/network/if-up.d/00check-network-cable script.
Story: 2009221
Task: 45758
Test Plan
PASS Build platform-kickstarts package
PASS Build ISO
PASS Run bootstap/unlock
PASS Check /etc/network/interface.d/ifcfg-* if 'sleep 20'
is still present.
Signed-off-by: Charles Short <charles.short@windriver.com>
Change-Id: I0d7a6387cb42a0cb1af62fa49d6790b641c78b75
The management interface is currently set to manual but should
dhcp for its IP address.
This update sets the management interface setup option to dhcp.
Test Plan:
PASS: Verify platform-kickstarts package build
PASS: Verify freshly installed system node goes online (hw)
PASS: Verify dhcp logs in daemon.log
Story: 2009968
Task: 45813
Signed-off-by: Eric MacDonald <eric.macdonald@windriver.com>
Change-Id: Ib25887a2cf16cc72af4dab090c62687c2254a2cb
