loop over branches

[jeremy mordkoff]

We have too many tests, so we have broken them up into three groups. One group that runs after each checkin and one group that runs once a day on any branch with changes and one group that I would like to limit to one instance at a time but still test any active branch. 

What is the best way to structure such a job? I am thinking a daemon type job that just iterates over branches and kicks off the real test job and then waited. The daemon would be sitting idle for almost 6 hours and then busy for 10 seconds, and then idle again, which seems slightly wasteful but at least this way jenkins would be self managing and I could tweak the daemon side to have priorities if I wanted, e.g. if master had changes and hadn't run in 36 hours, run it next, then go back to rotating fairly across the branches with changes. 

I can't have just one long running job that does it all because then my test results would get mixed together (I think) and I would not have red/yellow/blue for each branch. 

Any other ideas? 

--
You received this message because you are subscribed to the Google Groups "Jenkins Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-users/bcf0f35d-ed37-4ef6-8445-e628373c7817n%40googlegroups.com.

[Arnaud bourree]

Hello,

I've setup something similar.

We have one test platform and like you want to run tests on per branch.

I use lock resource plugin to run tests in exclusive stage.

Regards.

Arnaud 

Le mar. 1 déc. 2020 à 22:11, jeremy mordkoff <[hidden email]> a écrit :

We have too many tests, so we have broken them up into three groups. One group that runs after each checkin and one group that runs once a day on any branch with changes and one group that I would like to limit to one instance at a time but still test any active branch. 

What is the best way to structure such a job? I am thinking a daemon type job that just iterates over branches and kicks off the real test job and then waited. The daemon would be sitting idle for almost 6 hours and then busy for 10 seconds, and then idle again, which seems slightly wasteful but at least this way jenkins would be self managing and I could tweak the daemon side to have priorities if I wanted, e.g. if master had changes and hadn't run in 36 hours, run it next, then go back to rotating fairly across the branches with changes. 

I can't have just one long running job that does it all because then my test results would get mixed together (I think) and I would not have red/yellow/blue for each branch. 

Any other ideas? 

--
You received this message because you are subscribed to the Google Groups "Jenkins Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-users/bcf0f35d-ed37-4ef6-8445-e628373c7817n%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Jenkins Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-users/CAESUf_ArDwSr%3DjAHbYbQD0kR0RkOReDwds3gSAuJuoOBK57sXw%40mail.gmail.com.

[jeremy mordkoff]

I get that that is a better way of limiting the number of instances than using executor counts. Thanks for that. 

I still don't see a good way to schedule the jobs. Set a cron entry in the job so they all get scheduled once a day and then let all but one sit idle waiting on the resource lock? 

On Friday, December 4, 2020 at 2:27:55 PM UTC-5 Arnaud wrote:

Hello,

I've setup something similar.

We have one test platform and like you want to run tests on per branch.

I use lock resource plugin to run tests in exclusive stage.

Regards.

Arnaud 

Le mar. 1 déc. 2020 à 22:11, jeremy mordkoff <[hidden email]> a écrit :

We have too many tests, so we have broken them up into three groups. One group that runs after each checkin and one group that runs once a day on any branch with changes and one group that I would like to limit to one instance at a time but still test any active branch. 

What is the best way to structure such a job? I am thinking a daemon type job that just iterates over branches and kicks off the real test job and then waited. The daemon would be sitting idle for almost 6 hours and then busy for 10 seconds, and then idle again, which seems slightly wasteful but at least this way jenkins would be self managing and I could tweak the daemon side to have priorities if I wanted, e.g. if master had changes and hadn't run in 36 hours, run it next, then go back to rotating fairly across the branches with changes. 

I can't have just one long running job that does it all because then my test results would get mixed together (I think) and I would not have red/yellow/blue for each branch. 

Any other ideas? 

--
You received this message because you are subscribed to the Google Groups "Jenkins Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-users/bcf0f35d-ed37-4ef6-8445-e628373c7817n%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Jenkins Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-users/5adbc5e6-8a03-4d9b-b347-5582252c2641n%40googlegroups.com.