Reset RabbitMQ's net_ticktime to default 60 seconds
Bug #1598154 reported by
Dmitry Mescheryakov
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Fix Committed
|
High
|
Maksim Malchuk | ||
6.1.x |
Won't Fix
|
High
|
MOS Maintenance | ||
7.0.x |
Fix Released
|
High
|
Anton Chevychalov | ||
8.0.x |
Fix Released
|
High
|
Anton Chevychalov | ||
Mitaka |
Fix Released
|
High
|
Maksim Malchuk |
Bug Description
Right now we have net_ticktime set to 10 seconds in attempt to detect network partitions fast. The downside of that decision is that spurious network partitions/outages now trigger Erlang's partition handling logic.
The problem is, that partitioning handling logic is not bug-free and as a result, cluster might need to be fixed after short network hiccup. While if net_ticktime is 60 seconds, that would allow short outages to go unnoticed.
When selecting between 10 and 60 seconds net_ticktime, the 50 seconds difference in detection of partition could be neglected. At the same time, 10 seconds raises risk of downtime due to bugs in the logic. Hence default 60 seconds are preferable.
Changed in fuel: | |
status: | New → Confirmed |
tags: | added: area-mos |
tags: |
added: area-library removed: area-mos |
no longer affects: | fuel/newton |
Changed in fuel: | |
assignee: | MOS Oslo (mos-oslo) → Maksim Malchuk (mmalchuk) |
tags: | added: on-verification |
tags: | added: on-verification |
tags: | added: on-verification |
To post a comment you must log in.
Fix proposed to branch: master /review. openstack. org/337127
Review: https:/