[YARN-4470] [Umbrella] Application Master in-place upgrade

• Type: New Feature
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: resourcemanager
• Assignee: Giovanni Matteo Fumarola
• Reporter: Giovanni Matteo Fumarola
• Created: 18/Dec/15 01:57
• Updated: 23/Feb/16 22:29

Description

It would be nice if clients could ask for an AM in-place upgrade.
It will give to YARN the possibility to upgrade the AM, without losing the work
done within its containers. This allows to deploy bug-fixes and new versions of the AM incurring in long service downtimes.

Issue Links

• is part of: YARN-4726 [Umbrella] Allocation reuse for application upgrades
• relates to: YARN-3417 AM to be able to exit with a request saying "restart me with these (possibly updated) resource requirements"
• relates to: SLIDER-787 App Upgrade/Reconfig support in Slider

Sub-Tasks

• Add a public API to request an application upgrade: Open
• Introduce additional states in the app and app attempt state machines to keep track of the upgrade process: Open
• Add version information for the application and the application attempts: Open
• Implementation of the application master upgrade logic: Open
• Add the possibility to recover the state of an ongoing upgrade in case of RM restart: Open