x/kloudbuster
Code Issues Proposed changes

Update force_cleanup to handle properly volumes attached to non existent instances

Browse Source 
Change-Id: I262e2c612dc6f5f2318d186e866ebf4fcff8e440
This commit is contained in:
ahothan 2016-03-23 09:13:30 -07:00
parent 216772347f
commit b5b3628fce
2 changed files with 27 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
doc/source/cleanup.rst
View File

@ -43,6 +43,8 @@ that matches a given regular expression:
* users
* tenants
In some cases, because of timing reasons, you may have to run the force_cleanup script a few times
to get rid of all selected resources.
How to Select Resources to Delete
---------------------------------
@ -134,6 +136,20 @@ KloudBuster installation, the script requires the usual OpenStack python client
libraries and credentials.py (from the kloudbuster module). Otherwise, pick one
of the kloudbuster installation method to install the script.
Known Issues and Limitations
----------------------------
Volumes attached to instances that are no longer present cannot be deleted thorugh the Nova or Cinder APIs.
Such volumes will show up as attached to "None" and in the "in-use" or "available" state in the Horizon dashboard.
In this case, the script will print a warning with the volume ID:
WARNING: Volume 6080fdce-f894-4c41-9bc0-70120e8560a8 attached to an instance that no longer exists (will require manual cleanup of the database)
Cleanup of such volumes will require first setting the attach_status of the corresponding volume to "detached" in the Ciner database directly
(using a mysql cli such as "MariaDB [cinder]> UPDATE volumes SET attach_status='detached' WHERE id='18ed7f10-be49-4569-9e04-2fc4a654efee';")
then re-run the script (or manually delete the volume from Horizon).
Examples
--------

18
kloudbuster/force_cleanup.py
View File

@ -153,13 +153,17 @@ class StorageCleaner(AbstractCleaner):
vol = self.cinder.volumes.get(id)
if vol.attachments:
# detach the volume
if not self.dryrun:
ins_id = vol.attachments[0]['server_id']
self.nova.volumes.delete_server_volume(ins_id, id)
print ' . VOLUME ' + vol.name + ' detaching...'
else:
print ' . VOLUME ' + vol.name + ' to be detached...'
kb_detaching_volumes.append(vol)
try:
if not self.dryrun:
ins_id = vol.attachments[0]['server_id']
self.nova.volumes.delete_server_volume(ins_id, id)
print ' . VOLUME ' + vol.name + ' detaching...'
else:
print ' . VOLUME ' + vol.name + ' to be detached...'
kb_detaching_volumes.append(vol)
except NotFound:
print 'WARNING: Volume %s attached to an instance that no longer '\
'exists (will require manual cleanup of the database)' % (id)
else:
# no attachments
kb_volumes.append(vol)