During nova-powervc service startup, if the flavor sync
fails for some reasons, the code is blowing up causing
the service to get restarted abruptly. This fix is
to handle the same.
Change-Id: I3ce98153d0d3c0e184fd5ae44d7f0e54c19cebe2
Closes-Bug: #1819873
Signed-off-by: Arun Mani <arun.mani@in.ibm.com>
Update the if check in nova sync code to return instance
flavor from powervc, if the call to get the same from db
failed for some reasons. Also, update the keystone version
check in the common package.
Change-Id: Iafc489ad723ef6732bddda029f1976204deef37d
Closes-Bug: #1742377
boot instance with server name case insensitive, it should use
display_name which is case sensitive rather than hostname.
Change-Id: I742d83236ea3268fc64629e48c08f02562c7b11a
Closes-Bug: #1628404
Remove setting first scg id to boot parameter,
or this will fail in mult-scg environment.
Change-Id: I4e7349a2f5b6cbb6f59b629841442c9b8efe2fe4
Closes-Bug: #1608395
By default we use the host and available zone that ICM chooses, now I
will change it to the design that "Ignore the host that ICM selected,
defer it to powervc scheduler".
Change-Id: Ieaa53ebc425452ac5bad73441a32862373e456b9
Closes-Bug: #1538452
This is a test checkin to verify git process success that move the
repository from stackforge to openstack.
Change-Id: Ie79f79ce6428c517f20d8602ef8d1016ea9796fa
Closes-Bug: #1507455
List all resources across different service by adding "all-tenant"
when call restAPI
Change-Id: Ie29e0bac6fa4c3fe4b771489a722b6712edc3a28
Closes-Bug: #1496781
Catch quota reserve exception to avoid function break,
and catch get_info() excption to avoid startup break.
Change-Id: If2f8f535caeef81264f323df73927f0cfb4ecfb2
Closes-bug: #1493726
quotas are not updated when the volumes/instances are synced from
powervc side, this fix help to sync the quota during sync resources
from powervc side, increase quota when insert to local resource,
deduct quota when delete local resources.
Change-Id: Ic5a77e7d154b247d253db19d59ee66a7aa473995
Closes-Bug: 1493232
This bug tracks the implementation for volume sync and create/attach/detach
for multiattach function. Host status update also fixed along with this bug.
Change-Id: I45a776775473818c75388c3554664982849c4fae
Closes-Bug: #1491723
Quotas are not properly updated initially for on-boarded VMs from PowerVC,
when using powervc driver, need to commit quota after insert a pvc instance
when sync from powervc side.
Change-Id: I29916631129f9c3c1bb25b7fda947668328c57fd
Closes-Bug: #1472845
Block device mapping object mismatches with the latest community API,
which causes the sync problem.
Change-Id: I1e5059b1fb6f98e73b5fabbeb839a3281f8f29b5
Closes-Bug: #1455331
There is something wrong with volume v1 api to get volume detail info,
need to use volume v2 client instead of v1 to get local volume.
Change-Id: I76d067d2b1ee13c363dd5c22332d99d277c83f47
Closes-Bug: #1453358
Tempest failed due to openstack interface changed, fix it as replace all
unused driver method parameters with "*args, **kwargs"
Change-Id: I25c7e0f0eee7c8d16ac4de4e5918fa690544e33c
Closes-Bugs: #1451717
vm update failed due to nova api removed, need to use
instance object API instead
Change-Id: I5bdac5ec6b6af3340d3af3aff0cd4ae4859bd442
Closes-Bug: #1451131
The way of get flavor is change to get it from instance object,
need to update code to meet this change or driver sync up code
will fail. Also fix some other problem.
Change-Id: I8e79d7859b79e9d9736a1ba3c25ca56752cae6f8
Closes-Bug: #1432947
For Activation Engine, instance would have a configuration_data in its
system metadata, powervc driver does not need this and before save nova
instance, we need to remove this system metdata. Otherwise db will not
throw exception.
Closes-Bug: #1410649
Change-Id: I3fef4ae8c2d99dc103164fa346a2424185e773f5
When powervc driver found there is no or deleted image associated
during sync instances, it will set a default image which causes
exploiter crash, this patch will keep original image_ref when no
related image found.
Change-Id: I191d31cfdecaedd22b449728be732ca54057c5f9
Closes-Bug: 1404777
Consider after a while this "instance_type" has been changed to "flavor",
and we never use them, to avoid frequently update the powervc-driver
signature, append "**kwargs" to spawn() method.
Change-Id: I44c664c4e1650955e1eaa1f619223c00ade6f999
Closes-Bug: 1403378
1. Move the wait volume state logic from nova to cinder
2. Wait volume state to be "in-use", then complete volume attachment
3. Change volume sync code to not sync "attaching" or "detaching" state
as if sync code change such state, then volume attach/detach will be
broken
Closes-Bug: #1402467
Change-Id: I3640afe9ad4aba8788a932d05e94f55a3b8d6333
1. In powevc nova driver, it's better to use instance.save() instead of
db call
2. clear pvc_id after destory the powervc instance.
3. override pvc_id regardless powervc instance's metadata pvc_id.
Closes-Bug: #1395731
Change-Id: I2bb2494a8a2976064494101af811f6be22379472
1. Fix the previous unittest issue
2. Add test case to check no local port id found
Change-Id: I95674b45a6db64359a1d8a47bfc5aa367ab8594d
(cherry picked from commit 2919c4ef28)
PowerVC will set VM state to "Active" if the live migration failed as
the source host or dest host is running live migration. With our current
code, we are looping wait for the migration to be completed until the
powervc instance's host has been changed. That's definitely wrong..
Closes-Bug: #1393663
Change-Id: Ifb826bafa692b51f3d5997d427e10fbe17a21ca6
If the powervc instance is removed from driver's call or powervc side,
there will be a timing issue between the sync call and driver call, the
sync call still list the powervc instance before it insert the local
record. Add a global list to keep the removed powervc instance ids, and
avoid to create such local instance if the powervc instance has already
been deleted.
Closes-Bug: #13918752
Change-Id: If1e7993f4ae1517577c909fb71134ba50c88083f
nova manager will first delete local neutron port before call driver to
detach port on PowerVC. This behavior will trigger a local port deleted
event and cause current driver code failed to retrieve remote PowerVC
port id when doing interface detach
1. First priority is to use local port id to retrieve remote PowerVC
port id, which is the current logic.
2. If local port id cannot be retrieved, it will call interface_list on
remote PowerVC instance and get all attached interface information. Then
the code will scan all the local port IP addresses and compare to remote
IP addresses and find the matched port id to detach.
Change-Id: I3a87f3780c2cb949a0dc80380a37ab17bb98e059
Closes-Bug: #1392540
(cherry picked from commit 8c766d0861)
1. Add debug logs to sync code
2. fix an issue found the new added instance scg_id is None, there will
be an AttributeError thrown.
Closes-Bug: #1391875
Change-Id: Idaa6ec9982e915d8fec0e2e6d32d5bb5e49a674f
To make debug process conveniently , add more debug logs into
deploy/delete/resize methods , at the top of method and at the
end of the method , to make the process clearly , and print exception
stacktrace if necessary.
The log info should be like this :
1. Enter to $method instance of $instance_uuid
2. Exit to $method instance of $instance_uuid
Change-Id: I6678976c09ba923a0ada68a37a706642c18ed0f0
oslo.messaging is exposing a new API allow creating listener by
supplying a pool name, then the listeners with that pool name will use
the queue name as the pool name, multiple listeners can be listening
with the same topic.
Change-Id: I220d50fad8cda6f6f50f31215a7a6b98523c35dc
Closes-Bug: #1388663
For spawning, it will loop to check the powervc vm instance status,
there will be a timing window that the VM because active, then the sync
is running here, but the driver's spawning is till sleeping, then after
this sync code run, nova framework will fail with above exception..
This sync code need to be retained there as the VM might stick with
build/spawning state like the nova-powervc is abnormal killed...,
the fix solution can be:
1. create a dict in sync code, and put such state instance id to
this dict,
2. in next sync sycle, check if the instance is still in build/spawning,
but the powervc state is in acitve/none, then update status and remove
the id from the dict, this will delay the state updating in sync
process to avoid the time window that driver.spawning is sleeping
to poll the powervc instance state.
Change-Id: I7fe9f5f2b48c9758bc102ebf209d631c0308a316
Closes-Bug: 1388683