Applying MU#4 results in DHCP agents started on all controllers
Bug #1485174 reported by
Eugene Nikanorov
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Mirantis OpenStack |
Invalid
|
High
|
Eugene Nikanorov |
Bug Description
This results in tenant networks being scheduled on two DHCP agents, which is incompatible with dnsmasq settings provided by DHCP settings in 6.0.
Also, for some reason, external network got scheduled on one of DHCP agents too.
This particular issue might affect 7.0 too, although is minor.
In itself, scheduling networks to more than 1 agents (the setting in neutron.conf) is a bug.
But the issue is that the script should not start DHCP agent on controllers if it was not present, e.g. it should be more smart.
description: | updated |
Changed in mos: | |
milestone: | 6.0-updates → 6.0.1 |
To post a comment you must log in.
Eugene, please provide more information - details about the deployment (CentOS or Ubuntu, configuration), what were the steps to reproduce, was specific the error, what was expected behavior, thanks!