[JIRA] (OVIRT-1774) Upstream source collector can fail to push on
multi-branch projects
by Barak Korren (oVirt JIRA)
[ https://ovirt-jira.atlassian.net/browse/OVIRT-1774?page=com.atlassian.jir... ]
Barak Korren updated OVIRT-1774:
--------------------------------
Epic Link: (was: OVIRT-400)
> Upstream source collector can fail to push on multi-branch projects
> -------------------------------------------------------------------
>
> Key: OVIRT-1774
> URL: https://ovirt-jira.atlassian.net/browse/OVIRT-1774
> Project: oVirt - virtualization made easy
> Issue Type: Bug
> Components: pusher.py
> Reporter: Barak Korren
> Assignee: infra
> Labels: poll-upstream-sources, upstream-source-collector
>
> When the upstream source collection code looks for similar patches in order to avoid pushing a new patch, it lookes for patches that contain the same change to the '{{upstream-sources.yaml}}' file, regardless of the branch they may belong to.
> The code currently ignores the possibility that similar changes might be required for different branches, in which case different patches may need to be pushed.
> The way of detecting similar patches should be changed so that either the branch name is included in the checksum that are used to identify a patch, or the query is limited to patches of the branch being handled.
--
This message was sent by Atlassian Jira
(v1001.0.0-SNAPSHOT#100087)
6 years, 6 months
[JIRA] (OVIRT-1782) Make sure all VDSM build dependencies for FC27
are met.
by Barak Korren (oVirt JIRA)
[ https://ovirt-jira.atlassian.net/browse/OVIRT-1782?page=com.atlassian.jir... ]
Barak Korren updated OVIRT-1782:
--------------------------------
Resolution: Won't Fix
Status: Done (was: To Do)
VDSM development had already moved on to FC28
> Make sure all VDSM build dependencies for FC27 are met.
> -------------------------------------------------------
>
> Key: OVIRT-1782
> URL: https://ovirt-jira.atlassian.net/browse/OVIRT-1782
> Project: oVirt - virtualization made easy
> Issue Type: New Feature
> Components: CI client projects
> Reporter: Barak Korren
> Assignee: infra
> Labels: vdsm
>
> There had been attempts recently to make VDSM build on FC27.
> Those attempts were met with mixed success as issues kept surfacing.
> Fro now all the FC27 jobs for VDSM had been disabled.
> This ticket is for racking work of debugging issues to re-enable the jobs
--
This message was sent by Atlassian Jira
(v1001.0.0-SNAPSHOT#100087)
6 years, 6 months
[JIRA] (OVIRT-1831) Make `global-setup.sh` check if packages are
installed before trying to install them
by Barak Korren (oVirt JIRA)
[ https://ovirt-jira.atlassian.net/browse/OVIRT-1831?page=com.atlassian.jir... ]
Barak Korren reassigned OVIRT-1831:
-----------------------------------
Assignee: (was: infra)
> Make `global-setup.sh` check if packages are installed before trying to install them
> ------------------------------------------------------------------------------------
>
> Key: OVIRT-1831
> URL: https://ovirt-jira.atlassian.net/browse/OVIRT-1831
> Project: oVirt - virtualization made easy
> Issue Type: Improvement
> Components: Jenkins Slaves
> Reporter: Barak Korren
>
> The way '{{global-setup.sh}}' works currently is that it tries to install all the packages it needs in a single yum command and then runs in a loop and checks if the packages were installed properly.
> If the packages are already installed this takes a negligible amount of time on CentOS slaves, but on Fedora it seems to take much longer and as much as 2 minutes.
> We should try to make '{{global-setup.sh}}' faster by making it not try to install packages before checking if they are there, but instead check first and only try to install the missing ones.
> Note that if we'll install multiple packages in a single command, we'll still need to verify they were properly installed after the yum/dnf command finishes.
--
This message was sent by Atlassian Jira
(v1001.0.0-SNAPSHOT#100087)
6 years, 6 months
[JIRA] (OVIRT-1836) Optimize options normalization methods in
stdci_dsl
by Barak Korren (oVirt JIRA)
[ https://ovirt-jira.atlassian.net/browse/OVIRT-1836?page=com.atlassian.jir... ]
Barak Korren reassigned OVIRT-1836:
-----------------------------------
Assignee: (was: infra)
> Optimize options normalization methods in stdci_dsl
> ---------------------------------------------------
>
> Key: OVIRT-1836
> URL: https://ovirt-jira.atlassian.net/browse/OVIRT-1836
> Project: oVirt - virtualization made easy
> Issue Type: Improvement
> Components: STDCI DSL
> Reporter: Daniel Belenky
>
> Generally, we have 4 kinds of options: script, globals, options that affect host selection and options to configure the runner.
> We need to take care of global options first since they affect everything, then take care of script because it tells which vectors to throw away.
> Since runners and pipelines care about different sets of options, we need to split the options handling functions.
> This will enable us to compose much more efficient pipelines in the API.
--
This message was sent by Atlassian Jira
(v1001.0.0-SNAPSHOT#100087)
6 years, 6 months
[JIRA] (OVIRT-1831) Make `global-setup.sh` check if packages are
installed before trying to install them
by Barak Korren (oVirt JIRA)
[ https://ovirt-jira.atlassian.net/browse/OVIRT-1831?page=com.atlassian.jir... ]
Barak Korren updated OVIRT-1831:
--------------------------------
Epic Link: (was: OVIRT-400)
> Make `global-setup.sh` check if packages are installed before trying to install them
> ------------------------------------------------------------------------------------
>
> Key: OVIRT-1831
> URL: https://ovirt-jira.atlassian.net/browse/OVIRT-1831
> Project: oVirt - virtualization made easy
> Issue Type: Improvement
> Components: Jenkins Slaves
> Reporter: Barak Korren
> Assignee: infra
>
> The way '{{global-setup.sh}}' works currently is that it tries to install all the packages it needs in a single yum command and then runs in a loop and checks if the packages were installed properly.
> If the packages are already installed this takes a negligible amount of time on CentOS slaves, but on Fedora it seems to take much longer and as much as 2 minutes.
> We should try to make '{{global-setup.sh}}' faster by making it not try to install packages before checking if they are there, but instead check first and only try to install the missing ones.
> Note that if we'll install multiple packages in a single command, we'll still need to verify they were properly installed after the yum/dnf command finishes.
--
This message was sent by Atlassian Jira
(v1001.0.0-SNAPSHOT#100087)
6 years, 6 months
[JIRA] (OVIRT-1831) Make `global-setup.sh` check if packages are
installed before trying to install them
by Barak Korren (oVirt JIRA)
[ https://ovirt-jira.atlassian.net/browse/OVIRT-1831?page=com.atlassian.jir... ]
Barak Korren updated OVIRT-1831:
--------------------------------
Epic Link: (was: OVIRT-400)
> Make `global-setup.sh` check if packages are installed before trying to install them
> ------------------------------------------------------------------------------------
>
> Key: OVIRT-1831
> URL: https://ovirt-jira.atlassian.net/browse/OVIRT-1831
> Project: oVirt - virtualization made easy
> Issue Type: Improvement
> Components: Jenkins Slaves
> Reporter: Barak Korren
> Assignee: infra
>
> The way '{{global-setup.sh}}' works currently is that it tries to install all the packages it needs in a single yum command and then runs in a loop and checks if the packages were installed properly.
> If the packages are already installed this takes a negligible amount of time on CentOS slaves, but on Fedora it seems to take much longer and as much as 2 minutes.
> We should try to make '{{global-setup.sh}}' faster by making it not try to install packages before checking if they are there, but instead check first and only try to install the missing ones.
> Note that if we'll install multiple packages in a single command, we'll still need to verify they were properly installed after the yum/dnf command finishes.
--
This message was sent by Atlassian Jira
(v1001.0.0-SNAPSHOT#100087)
6 years, 6 months
[JIRA] (OVIRT-1836) Optimize options normalization methods in
stdci_dsl
by Barak Korren (oVirt JIRA)
[ https://ovirt-jira.atlassian.net/browse/OVIRT-1836?page=com.atlassian.jir... ]
Barak Korren updated OVIRT-1836:
--------------------------------
Epic Link: (was: OVIRT-400)
> Optimize options normalization methods in stdci_dsl
> ---------------------------------------------------
>
> Key: OVIRT-1836
> URL: https://ovirt-jira.atlassian.net/browse/OVIRT-1836
> Project: oVirt - virtualization made easy
> Issue Type: Improvement
> Components: STDCI DSL
> Reporter: Daniel Belenky
> Assignee: infra
>
> Generally, we have 4 kinds of options: script, globals, options that affect host selection and options to configure the runner.
> We need to take care of global options first since they affect everything, then take care of script because it tells which vectors to throw away.
> Since runners and pipelines care about different sets of options, we need to split the options handling functions.
> This will enable us to compose much more efficient pipelines in the API.
--
This message was sent by Atlassian Jira
(v1001.0.0-SNAPSHOT#100087)
6 years, 6 months
[JIRA] (OVIRT-1836) Optimize options normalization methods in
stdci_dsl
by Barak Korren (oVirt JIRA)
[ https://ovirt-jira.atlassian.net/browse/OVIRT-1836?page=com.atlassian.jir... ]
Barak Korren updated OVIRT-1836:
--------------------------------
Epic Link: (was: OVIRT-400)
> Optimize options normalization methods in stdci_dsl
> ---------------------------------------------------
>
> Key: OVIRT-1836
> URL: https://ovirt-jira.atlassian.net/browse/OVIRT-1836
> Project: oVirt - virtualization made easy
> Issue Type: Improvement
> Components: STDCI DSL
> Reporter: Daniel Belenky
> Assignee: infra
>
> Generally, we have 4 kinds of options: script, globals, options that affect host selection and options to configure the runner.
> We need to take care of global options first since they affect everything, then take care of script because it tells which vectors to throw away.
> Since runners and pipelines care about different sets of options, we need to split the options handling functions.
> This will enable us to compose much more efficient pipelines in the API.
--
This message was sent by Atlassian Jira
(v1001.0.0-SNAPSHOT#100087)
6 years, 6 months