older version of pegleg render command had -o option to save the renedered yaml
with latest version of pegleg render command, -s is used to save renedered yaml
Change-Id: Ia4312a6a17ba88c86cea215163ce732bb06e9d37
Setting PEGLEG_PASSPHRASE and PEGLEG_SALT to pegleg container
to use it for encryption/decryption.
Change-Id: I2c5c01fb0ef7366dbc35d4f8730ac1aa9ca63db1
This PS removes set -x from general runs so that all the commands
will not be thrown in output.
Change-Id: I8068e170d632518a93f5bf097d3a88cc3af01433
Signed-off-by: Sreejith Punnapuzha <sreejith.punnapuzha@outlook.com>
This ps changes set -x to set -e in config.sh so that all the
commands will not throw in output
Change-Id: I08d2e56ea300509a3f579c888bfe6f1bd69deefd
Signed-off-by: Sreejith Punnapuzha <sreejith.punnapuzha@outlook.com>
This PS replaces resolv.conf link file with generated resolv.conf
and stops systemd-resolved service on build node which causes issue
with coredns container running on build node
Change-Id: Id317587b0c46fdc4822ac9b5bb27fef90bb17636
Signed-off-by: Sreejith Punnapuzha <sreejith.punnapuzha@outlook.com>
This PS enables vnc console for vms created by aiab scripts.
Change-Id: I7e882b80986a7a0868b793ab9783098cfc09092a
Signed-off-by: Sreejith Punnapuzha <sreejith.punnapuzha@outlook.com>
Make files collected and rendered by Pegleg (owned by root
and 640 by default) world-readable so non-root Promenade
processes can read them.
Change-Id: Ic5bce645ebf58c106fd59701b3f282f9e2455f0e
After timeout pipeline is trying to run deploy_site again and failed.
There is no info in debug report and console output about current status
if the shipyard action. Added new cmd to describe the action.
ceph-mon is detecting low disk space on n0, it's <70%. So, small increase
of disk space is needed.
Change-Id: Ia92390ccadec4485c37038187034a2978911accd
- Move the default temp location to /var/tmp so that gate
workspace directories persist across reboots
Change-Id: Ic11d163c3aba36a08f42bd1739003446aed3a0ce
kubectl_cmd uses ssh_cmd to run kubectl over ssh. ssh_cmd does not
allow to pass extra env vars in command.
So updated to use ssh_cmd_raw with env KUBECONFIG set
Change-Id: I1f2391a79002db15785644e10e673936c3dff4ed
PODs are being evicted due reduced disk size (30 from 64G).
The issue is constantly appearing around MAAS image deployment.
Change-Id: If8a0717a7f57d77d0c91a6798c774dc9cc06ecb6
- Add a shellcheck linter for the scripts in the multinode
framework
- Update all scripting to comply with shellcheck
- Move linting job to Ubuntu Bionic as the multinode gate now
requires Bionic versions of libvirt
Change-Id: Ibee645331421e1e6cecd4e3daa8e9c321dce5523
- With a extensible network configuration, selecting
the correct IP for a node needs a little more intelligence
Change-Id: I742270a4df440c02bb4bcc03badae3a96e094221
- Start documentation on crafting a framework scenario
manifest now that the framework supports significant
manifest-driven customization.
Change-Id: I7ed051238ce9262641615103ec73af3b0b1cc630
- In some cases, a user may want to provide their own stage
library or replace one of these built-in stages. Allow external
(to this framework) directories be specified as containing
stage scripts.
Change-Id: I468ea56a45e3c041e10040433e70eb9aa354ad9a
This adds scripts that gather cluster object information
and namespaces object information. These scripts then
create a folder and both yaml and txt files for each
object
Change-Id: Ia22caef4503451e637b20e1d62c4bd50aedfece2
Prior change[0] breaks if apparmor profiles are not present in the
manifest. This commit allows the script to proceed if they are not
found.
[0]: https://review.opendev.org/#/c/676532/
Change-Id: Idc3f458c0002c707afceb6609de6822fb638f608
The bootactions which will be deployed via Drydock on nodes need
to be performed on Genesis node as well. This should be done as
part of pre-genesis setup before genesis.sh is executed.
This patchset allows apparmor profiles to be staged on the genesis
node in the same manner as seccomp profiles.
Change-Id: I418c955a131dc8e23ab5f4d900ae5ea4f7985468
Reference: https://review.opendev.org/644824
When running kubectl command from ssh_cmd, the env variable KUBECONFIG is not visible
and so the kubectl commands are failing with below error
"The connection to the server localhost:8080 was refused - did you specify the right host or port?"
Change-Id: I453f95b745a914aad5c608c5e5f625056e516add
- The persistence of the br_netfilter disable settings wasn't
working due to the br_netfilter kmod loading after sysctl
settings are applied. Add a udev rule so that the sysctl
settings are applied when the module loads.
Change-Id: I31eae66f953e644c09b86d5449ac79cf253d5df3
- Create the besteffort utility function to run a command
outside of set -e
- Use besteffort to disable brnetfilter so the gate survives
when the kernel module isn't loaded.
Change-Id: Ic79aff8cce357d02715a305664ca700e54822627
Prepare hyperkube binary before running promenade container.
This is to simplify procedure when using hyperkube image.
Change-Id: I04dfdef1247e7d1701e673e53132e6a9208d3c64
This change adds a script that enables the ability to add a site test
stage to a manifest. The Shipyard test site action runs Helm tests for a
deployed site.
Change-Id: I77c123c89fe1e780ede899fa3cc6d2a726f74f49
New source is Docker image that can be used to exctract hyperkube binary.
It can be done by running Docker container inside Promenade container
and copy binary to mounted volume.
Change-Id: Ie79ff4b349e6321e7a335359b2c461dac783b431
This PS makes bridge-nf configuration persisten as part of
setup_gate.sh
Change-Id: Ic8a4f231d277716152dfe029e1961f165a7dec11
Signed-off-by: Sreejith Punnapuzha <Sreejith.Punnapuzha@outlook.com>