story: 2003389
Task: 24556
It is removed when upgrade rpm package to CentOS 7.5 version.
Add it back to keep it the same as before.
For rdma-core package, it could be generated by src rpm, so no need to have rpm package here.
Change-Id: Ia6cced5362e9a844ecaa92e40c8f913f2206c337
Signed-off-by: slin14 <shuicheng.lin@intel.com>
solve "Error: no symbol table" issue
This commit fixes:
The system show "“Error: no symbol table.
Press any key to continue…” when booting
Closes-Bug: 1792257
Change-Id: I49d78bc6b5c23da914a2095737959cf05dccc6c0
Signed-off-by: Sun Austin <austin.sun@intel.com>
Add sanity around the use of 'rm' in script. Make sure the file to be
deleted is the correct type, and that its content also meets
expectation.
An example of dangerous outcome in this code from using "sudo rm -rf" is
that if the CONTROLLER or COMPUTE variables are preceded by a space
character then the directory /var/lib/libvirt/images/ will be deleted.
Refrain from using recursive and force options, and in this case because
they are not needed.
Related-Bug: #1790716
Change-Id: I76797133589d993dca2b2aac3c97184bac0457ca
Signed-off-by: Michel Thebeau <michel.thebeau@windriver.com>
Story: 2003596
Task: 24917
kernel version is upgraded from 862.6.3 to 862.11.6
Change-Id: If2efecc9510617cae5645ca07732620a40b2ebed
Signed-off-by: slin14 <shuicheng.lin@intel.com>
In the event that someone clones the stx-tools repo without
fore-knowledge of the wiki, this is handy. The author prepares this
commit from the experience doing just that.
Related-Bug: #1790716
Change-Id: If3d1e250e1c2c246ac894117fc649d6f145f22f1
Signed-off-by: Michel Thebeau <michel.thebeau@windriver.com>
The INTERNAL and EXTERNAL networks configured by destroy and setup
network scripts are inverted with reference to the default
config_controller selections as well as the wiki
StarlingX/Installation_Guide_Virtual_Environment/Controller_Storage.
The most pressing concern in this is that what was actually the internal
management network was being nat'd.
Also, under normal testing it is not recommended to address the internal
management network from the virtualization host. There are unlikely to be
configurations that will have external devices addressing the management
network.
Delete what was "EXTERNAL", actually internal, and rename "INTERNAL" to
be EXTERNAL.
Related-Bug: #1790716
Change-Id: I08f9e7712fed120001c864903880e13e9478764d
Signed-off-by: Michel Thebeau <michel.thebeau@windriver.com>
Tests should run on a comparably newer model of CPU and should be fixed
for all tests under virtualization.
Nehalem cpu model is recommended as one that has been used for this
purpose for a few years.
Specifying 'forbid' for fallback and 'exact' for mode will ensure that
hosts the Qemu emulator is up-to-date on the host, capable of emulating
the expect model. The consistency will make test results reliable
across host machines.
The newer machine type supports ACPI for example.
The feature "vmx" is required on computes to support nesting.
Related-Bug: #1790716
Change-Id: Ib338501b4805532a9d1ebaf55735f9084bfa8e96
Signed-off-by: Michel Thebeau <michel.thebeau@windriver.com>
Some of the entries are invalid when changing the machine type, which is the
next commit to follow this one.
Libvirt will fill these things in correctly for the specified hardware.
Related-Bug: #1790716
Change-Id: I674657cca117563de0c0ff04231224198623d3dd
Signed-off-by: Michel Thebeau <michel.thebeau@windriver.com>
The xml template is used to define both controllers, and so the two will
have the same macs.
Libvirt will fill in the omitted mac addresses.
Related-Bug: #1790716
Change-Id: Ib9544b5b8cbc2e5cf8ea941129ee9720cba3df97
Signed-off-by: Michel Thebeau <michel.thebeau@windriver.com>
Story: 2003340
Task: 24378
No patch is needed for this package, so replace it with rpm package
Change-Id: I21eecb9fb898cf0c7484eb03376fc4703c81204e
Signed-off-by: slin14 <shuicheng.lin@intel.com>
Customization to memcached package is removed so no need to download
source rpm anymore. Download the binary rpm instead.
Story: 2002826
Task: 26346
Change-Id: I3acd809c81556350461f4e017f91b00a3bb97473
Signed-off-by: Jack Ding <jack.ding@windriver.com>
This reverts commit 9d5971d602.
We have found root cause for config_controller failing at step 6 and fix it.
Story: 2003389
Task: 24504
Change-Id: I7fdc1b81a96ac1c26fd8aa2a4754c021499cd3a9
The dl_rpms.sh script may be passed an optional yum.conf file. If
this file is used, then the releasever variable is read from the
rather than use the default releasever=7. If the releasever is
not defined in the file, the grep which searches for the variable
will fail, and the dl_rpms.sh script exits with an error.
This change checks for the presence of releasever before attempting
to read it from the file. Change made to dl_rpms.sh and
mirror-check.sh
Tested using yum.conf files both with and without releasever
present.
Change-Id: Ie13241e5662188d75b5d147537915c0a35ea7397
Closes-Bug: 1791113
Signed-off-by: Jason McKenna <jason.mckenna@windriver.com>
Story: 2003389
Task: 24556
Some rpms are not upgraded for different reason:
1. libguestfs/python-libguestfs is not upgraded due to 7.5 version needs libvirt-daemon-kvm >= 3.9.0-1
2. puppet modules rpm is not upgraded due to openstack puppet has dependency check.
3. packages has higher version than CentOS7.5 is not changed.
Change-Id: I9e9dee9a51cdbed9d486dd802725b0956bfe4a3f
Signed-off-by: slin14 <shuicheng.lin@intel.com>
Also:
* Use build-openstack-docs-pti job template for docs
* Use build-openstack-releasenotes job for release notes
(We can't use the OpenStack releasenotes template as it includes
publish jobs, stx needs its own)
* Add newnote tox environment as convenience for creating new release
notes, re-using the releasenotes venv.
* Create a release summary note.
* Remove the unused 'true' tox environment.
* Fix doc/build in .gitignore
Change-Id: Ibe9d298bcc76f023c0500038db55d8384f5ff731
Signed-off-by: Dean Troyer <dtroyer@gmail.com>
This changes branch-stx.sh to not assume that a manifest file will
be used to create the list of repos to process, and allow direct
repo URLs on the command line.
Change-Id: Ie2bbfa314b33d2b2b810ee479b4a0158e9d819bf
Signed-off-by: Dean Troyer <dtroyer@gmail.com>
This solves a problem when the default releasever is not set for
distros that dont have a YUM as a standard tool.
Read from the yum.conf if set on the command line since that is mostly
likely going to be when we need to know this, in the default case (no -c
yum.conf) it will get the correct value from the system.
Change-Id: I21b0e8f881d054d8cdf6cdd831836a7f3e3020b2
Signed-off-by: Saul Wold <sgw@linux.intel.com>
This updated package will cause config_controller failed at step 6.
This reverts commit 454b9d6d75.
Change-Id: I71dcecd63527223589123eb6dcfa151755a3bd60
Eight log files per RPM list is complicated for error detection and
managing. This commmit is changing that to three logs.
One for general information about errors, other for missing RPMs and
another for missing SRPMs. These could be used by download_mirror.sh
script to look for packages with K1.
All logs are now kept in logs directory.
And now the download_mirror.sh has a function for generating logs'
names.
Change-Id: I901657bc56209d54a3fe5be492d76716249f2755
Signed-off-by: Marcela Rosales <marcela.a.rosales.jimenez@intel.com>
* Split the network setup/teardown bits into their own scripts
so you can shoot yourself in the foot with intention
rather then unexpectedly.
* Cowardly refuse to configure a network if the first interface name exists
* Change the default bridge name to stxbr
* Make the network variables all configurable via the environment
* Don't make assumptions about where ISOIMAGE is located
* Include a basic doc that outlines the differences from existing
installation steps. (Needs more!)
Change-Id: Ic46c03a09da97765b9f6bfe07e089efa38738993
Signed-off-by: Dean Troyer <dtroyer@gmail.com>
This adds two sets of deployments scripts to aid in setting up virtual
environments for testing StarligX, for libvirt/qemu and VirtualBox.
This is the first exposure of an internal project and is not fully
error-proofed, it should be examined carefully to understand what
the scripts will do to your system before running them.
The first set of fixes for this to make it less instrusive and work
in a shared server follow in https://review.openstack.org/#/c/597643.
There are also a couple of bashate errors fixed since that is now a voting job...
[NOTE(dtroyer): I converted the original commit message into a README
because that's information that needs to be with the scripts.]
Needed-by: https://review.openstack.org/#/c/597643
Change-Id: I0a6a148720b7a239380fd48f7ffdab272472e664
Co-authored-by: Lianhao Lu <lianhao.lu@intel.com>
Co-authored-by: Brian Avery <brian.avery@intel.com>
Co-authored-by: Nakul Dahiwade <nakul.dahiwade@intel.com>
Co-authored-by: Yan Chen <yan.chen@intel.com>
Co-authored-by: Ruijing Guo<ruijing.guo@intel.com>
Co-authored-by: Shuicheng Lin <shuicheng.lin@intel.com>
Co-authored-by: Felipe de Jesus Ruiz Garcia <felipe.de.jesus.ruiz.garcia@intel.com>
Co-authored-by: Jose Perez Carranza <jose.perez.carranza@intel.com>
Co-authored-by: Eddie Ramirez <eddie.ramirez@intel.com>
Co-authored-by: Kailun Qin <kailun.qin@intel.com>
Co-authored-by: Forrest Zhao <forrest.zhao@intel.com>
Co-authored-by: Kailun Qin <kailun.qin@intel.com>
Signed-off-by: Abraham Arce <abraham.arce.moreno@intel.com>
Signed-off-by: Dean Troyer <dtroyer@gmail.com>
OpenDev Sysadmins