deterministic hash for partitioning

Bug #1745184 reported by gordon chung
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Ceilometer
New
Undecided
Unassigned

Bug Description

in notification agent, we use hash() to decide which queue to shard data to. this might not return consistent results which could cause related metrics/events to be push to different nodes. if this happens, our transformers (and batching) get screwed up.

switch away from hash()

https://github.com/openstack/ceilometer/blob/6d27648c2983c6100322c5dfa4057f7255619c96/ceilometer/pipeline/base.py#L74

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.