[MESOS-3078] Recovered resources are not re-allocated until the next allocation delay.

• Type: Improvement
• Status: Reviewable
• Priority: Major
• Resolution: Unresolved
• Component/s: allocation
• Epic Link: allocator performance phase 2
• Story Points: 5
• Assignee: Guangya Liu
• Reporter: Benjamin Mahler
• Shepherd: Benjamin Mahler
• Created: 17/Jul/15 22:36
• Updated: 09/Feb/18 01:49

Description

Currently, when resources are recovered, we do not perform an allocation for that slave. Rather, we wait until the next allocation interval.

For small task, high throughput frameworks, this can have a significant impact on overall throughput, see the following thread:
http://markmail.org/thread/y6mzfwzlurv6nik3

We should consider immediately performing a re-allocation for the slave upon resource recovery.

Issue Links

• is blocked by: MESOS-4766 Improve allocator performance.
• is related to: MESOS-4811 Reusable/Cacheable Offer
• is related to: MESOS-4102 Quota doesn't allocate resources on slave joining.
• is related to: MESOS-6904 Perform batching of allocations to reduce allocator queue backlogging.
• is related to: MESOS-3157 Only perform periodic resource allocations.
• is related to: MESOS-4553 Manage offers in allocator.
• relates to: MESOS-4302 Offer filter timeouts are ignored if the allocator is slow or backlogged.