From b5ae98df9ea1d4f3c064fa852b1c83d84a415a77 Mon Sep 17 00:00:00 2001 From: Matt Riedemann Date: Wed, 23 Oct 2019 11:04:50 -0400 Subject: [PATCH] doc: note the need to configure cinder auth in reclaim_instance_interval Without configuring the [cinder] auth options in nova.conf, reclaiming SOFT_DELETED instances will fail to cleanup attached volumes since there is no user token in a periodic task and we have to rely on config to use a service user for that cleanup with cinder. This adds a note to the reclaim_instance_interval config option help to explain that. Change-Id: I9da30b53b4d1bcaca7944700a67a19523f3bcf5d Closes-Bug: #1849481 --- nova/conf/compute.py | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/nova/conf/compute.py b/nova/conf/compute.py index fccebabb37..0a8481a8fb 100644 --- a/nova/conf/compute.py +++ b/nova/conf/compute.py @@ -1042,11 +1042,23 @@ must be set globally otherwise servers could be put into a soft deleted state in the API and never actually reclaimed (deleted) on the compute node. +.. note:: When using this option, you should also configure the ``[cinder]`` + auth options, e.g. ``auth_type``, ``auth_url``, ``username``, etc. + Since the reclaim happens in a periodic task, there is no user token + to cleanup volumes attached to any SOFT_DELETED servers so nova must + be configured with administrator role access to cleanup those + resources in cinder. + Possible values: * Any positive integer(in seconds) greater than 0 will enable this option. * Any value <=0 will disable the option. + +Related options: + +* [cinder] auth options for cleaning up volumes attached to servers during + the reclaim process """), cfg.IntOpt('volume_usage_poll_interval', default=0,