transport_url is not updated nova_api.cell_mappings table, causi oslo.messaging._drivers.impl_rabbit EHOSTUNREACH,

Bug #1943864 reported by Mirek
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
OpenStack Nova Cloud Controller Charm
New
Undecided
Unassigned

Bug Description

Hi,

It looks like if you change any rabbitmq details like port or IP, e.g. enable SSL or rabbitmq redeployment it won't update nova_api.cell_mappings table with correct details and cause
2021-09-16 13:59:09.475 42555 ERROR oslo.messaging._drivers.impl_rabbit [req-563af0a9-3f21-428f-be53-630eda324615 - - - - -] Connection failed: [Errno 113] EHOSTUNREACH (retrying in 32.0 seconds): OSError: [Errno 113] EHOSTUNREACH

Which will cause OpenStack to hang on scheduling stage as nova-cc is trying to use a non-exisitng rabbitmq host.

I haven't got enviroment to re-produce it but it should as easy as once deploy, change rabbit to SSL or just remove and redeploy rabbitmq.

What fixed it in my case was to manually update the table. e.g.
update nova_api.cell_mappings
set transport_url = 'rabbit://nova:<pass>@<IP>:5672,nova:<pass>@<IP>:5672,nova:<pass>@<IP>:5672/openstack'
where id = 2;

According to the documentation, Nova will use transport_url from DB first and fall back to nova.conf only if not found in DB.
https://docs.openstack.org/nova/latest/user/cells.html
"Note
If you don’t specify the database and transport urls then nova-manage will use the [database]/connection and [DEFAULT]/transport_url values from the config file.
"

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.