allow multiple on demand requests to be queued up #950

Merged
merged 2 commits into from Apr 28, 2016

Conversation

Projects
None yet
2 participants
@ssalinas
Member

ssalinas commented Mar 14, 2016

@ssalinas ssalinas added the hs_staging label Mar 14, 2016

@ssalinas

This comment has been minimized.

Show comment
Hide comment
@ssalinas

ssalinas Mar 14, 2016

Member

@wsorenson @tpetr traced back through the different PendingTypes we could possibly get in combination with the ONEOFF type. Those are: RETRY, DECOMISSIONED_SLAVE_OR_RACK, UPDATED_REQUEST, or TASK_DONE. No others are currently created for an on demand request.

RETRY -> create another identical oneoff, there is no conflict here as both kinds just schedule a single task with the relevant options
DECOMISSIONED_SLAVE_OR_RACK - > no problem there, no new tasks will be scheduled for a oneoff
TASK_DONE -> Similar, we do not schedule new tasks for a oneoff on task done
UPDATED_REQUEST -> This was the only possible relevant one (for values like killOldNonLongRunningTasksAfterMillis which I don't believe we wired up for on demand anyway). However, when we encounter this pending type the update to request data is already finished and we do not schedule any new oneoff tasks for a UPDATED_REQUEST pending type

Member

ssalinas commented Mar 14, 2016

@wsorenson @tpetr traced back through the different PendingTypes we could possibly get in combination with the ONEOFF type. Those are: RETRY, DECOMISSIONED_SLAVE_OR_RACK, UPDATED_REQUEST, or TASK_DONE. No others are currently created for an on demand request.

RETRY -> create another identical oneoff, there is no conflict here as both kinds just schedule a single task with the relevant options
DECOMISSIONED_SLAVE_OR_RACK - > no problem there, no new tasks will be scheduled for a oneoff
TASK_DONE -> Similar, we do not schedule new tasks for a oneoff on task done
UPDATED_REQUEST -> This was the only possible relevant one (for values like killOldNonLongRunningTasksAfterMillis which I don't believe we wired up for on demand anyway). However, when we encounter this pending type the update to request data is already finished and we do not schedule any new oneoff tasks for a UPDATED_REQUEST pending type

@ssalinas

This comment has been minimized.

Show comment
Hide comment
@ssalinas

ssalinas Mar 14, 2016

Member

quick addition, misread an if statement, we in fact do not get a UPDATED_REQUEST pending type for on demand requests

Member

ssalinas commented Mar 14, 2016

quick addition, misread an if statement, we in fact do not get a UPDATED_REQUEST pending type for on demand requests

@jhaber

This comment has been minimized.

Show comment
Hide comment
@jhaber

jhaber Mar 14, 2016

Member

There are no words to express my excitement 👯

Member

jhaber commented Mar 14, 2016

There are no words to express my excitement 👯

@ssalinas ssalinas added the hs_qa label Mar 16, 2016

@ssalinas ssalinas modified the milestone: 0.4.12 Mar 18, 2016

@ssalinas ssalinas modified the milestones: 0.5.0, 0.6.0 Apr 5, 2016

@ssalinas ssalinas added the hs_stable label Apr 6, 2016

@ssalinas

This comment has been minimized.

Show comment
Hide comment
@ssalinas

ssalinas Apr 28, 2016

Member

This has been performing well with no issues, going to merge

Member

ssalinas commented Apr 28, 2016

This has been performing well with no issues, going to merge

@ssalinas ssalinas merged commit f257530 into master Apr 28, 2016

2 checks passed

continuous-integration/travis-ci/pr The Travis CI build passed
Details
continuous-integration/travis-ci/push The Travis CI build passed
Details

@ssalinas ssalinas deleted the on_demand_multi branch Apr 28, 2016

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment