[Yahoo-eng-team] [Bug 1382630] [NEW] access_ip_* not updated on reschedules when they should be
Public bug reported: For virt drivers that require networks to be reallocated on nova reschedules, the access_ip_v[4|6] fields on Instance are not updated. This bug was introduced when the new build_instances path was added. This new path updates access_ip_* before the instance goes ACTIVE... and it only updates when its not already set. The old path only updated the access_ip_* fields when the instance went ACTIVE... ** Affects: nova Importance: Undecided Status: New -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1382630 Title: access_ip_* not updated on reschedules when they should be Status in OpenStack Compute (Nova): New Bug description: For virt drivers that require networks to be reallocated on nova reschedules, the access_ip_v[4|6] fields on Instance are not updated. This bug was introduced when the new build_instances path was added. This new path updates access_ip_* before the instance goes ACTIVE... and it only updates when its not already set. The old path only updated the access_ip_* fields when the instance went ACTIVE... To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1382630/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1342919] Re: instances rescheduled after building network info do not update the MAC
I'm proposing a new virt driver method for nova that allows the virt driver to say whether reschedules should deallocate networks. Once the nova side is confirmed, we'll add the method to ironic's virt driver. ** Also affects: ironic Importance: Undecided Status: New ** Changed in: nova Assignee: Robert Collins (lifeless) => Chris Behrens (cbehrens) ** Changed in: ironic Assignee: (unassigned) => Chris Behrens (cbehrens) ** Changed in: ironic Status: New => In Progress -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1342919 Title: instances rescheduled after building network info do not update the MAC Status in OpenStack Bare Metal Provisioning Service (Ironic): In Progress Status in OpenStack Compute (Nova): In Progress Bug description: This is weird - Ironic has used the mac from a different node (which quite naturally leads to failures to boot!) nova list | grep spawn | 6c364f0f-d4a0-44eb-ae37-e012bbdd368c | ci-overcloud-NovaCompute3-zmkjp5aa6vgf | BUILD | spawning | NOSTATE | ctlplane=10.10.16.137 | nova show 6c364f0f-d4a0-44eb-ae37-e012bbdd368c | grep hyperv | OS-EXT-SRV-ATTR:hypervisor_hostname | b07295ee-1c09-484c-9447-10b9efee340c | neutron port-list | grep 137 | 272f2413-0309-4e8b-9a6d-9cb6fdbe978d || 78:e7:d1:23:90:0d | {"subnet_id": "a6ddb35e-305e-40f1-9450-7befc8e1af47", "ip_address": "10.10.16.137"} | ironic node-show b07295ee-1c09-484c-9447-10b9efee340c | grep wait | provision_state| wait call-back | ironic port-list | grep 78:e7:d1:23:90:0d # from neutron | 33ab97c0-3de9-458a-afb7-8252a981b37a | 78:e7:d1:23:90:0d | ironic port-show 33ab97c0-3de9-458a-afb7-8252a981 ++---+ | Property | Value | ++---+ | node_uuid | 69dc8c40-dd79-4ed6-83a9-374dcb18c39b | # Ruh-roh, wrong node! | uuid | 33ab97c0-3de9-458a-afb7-8252a981b37a | | extra | {u'vif_port_id': u'aad5ee6b-52a3-4f8b-8029-7b8f40e7b54e'} | | created_at | 2014-07-08T23:09:16+00:00 | | updated_at | 2014-07-16T01:23:23+00:00 | | address| 78:e7:d1:23:90:0d | ++---+ ironic port-list | grep 78:e7:d1:23:9b:1d # This is the MAC my hardware list says the node should have | caba5b36-f518-43f2-84ed-0bc516cc89df | 78:e7:d1:23:9b:1d | # ironic port-show caba5b36-f518-43f2-84ed-0bc516cc ++---+ | Property | Value | ++---+ | node_uuid | b07295ee-1c09-484c-9447-10b9efee340c | # and tada right node | uuid | caba5b36-f518-43f2-84ed-0bc516cc89df | | extra | {u'vif_port_id': u'272f2413-0309-4e8b-9a6d-9cb6fdbe978d'} | | created_at | 2014-07-08T23:08:26+00:00 | | updated_at | 2014-07-16T19:07:56+00:00 | | address| 78:e7:d1:23:9b:1d | ++---+ To manage notifications about this bug go to: https://bugs.launchpad.net/ironic/+bug/1342919/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1352510] [NEW] Delete and re-add of same node to compute_nodes table is broken
Public bug reported: When a compute node is deleted (or marked deleted) in the DB and another compute node is re-added with the same name, things break. This is because the resource tracker caches the compute node object/dict and uses the 'id' to update the record. When this happens, rt.update_available_resources will raise a ComputeHostNotFound. This ends up short-circuiting the full run of the update_available_resource() periodic task. This mostly applies when using a virt driver where a nova-compute manages more than 1 "hypervisor". ** Affects: nova Importance: Medium Assignee: Chris Behrens (cbehrens) Status: In Progress ** Changed in: nova Status: New => In Progress ** Changed in: nova Assignee: (unassigned) => Chris Behrens (cbehrens) ** Changed in: nova Importance: Undecided => Medium -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1352510 Title: Delete and re-add of same node to compute_nodes table is broken Status in OpenStack Compute (Nova): In Progress Bug description: When a compute node is deleted (or marked deleted) in the DB and another compute node is re-added with the same name, things break. This is because the resource tracker caches the compute node object/dict and uses the 'id' to update the record. When this happens, rt.update_available_resources will raise a ComputeHostNotFound. This ends up short-circuiting the full run of the update_available_resource() periodic task. This mostly applies when using a virt driver where a nova-compute manages more than 1 "hypervisor". To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1352510/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1347778] Re: raising Maximum number of ports exceeded is wrong
** Also affects: python-neutronclient Importance: Undecided Status: New -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1347778 Title: raising Maximum number of ports exceeded is wrong Status in OpenStack Compute (Nova): Triaged Status in Python client library for Neutron: Confirmed Bug description: When neutron API in nova calls create_port(), it looks for exceptions. Any 409 is turned into 'Maximum number of ports exceeded'. This is a horrible assumption. Neutron can return 409s for more than just this reason. Another case where neutron returns a 409 is this: 2014-07-22 18:10:27.583 26577 INFO neutron.api.v2.resource [req-b7267ae5-bafa-4c34-8e25-9c0fca96ad2d None] create failed (client error): Unable to complete operation for network ----. The mac address XX:XX:XX:XX:XX:XX is in use. This can occur when the request to create a port includes the mac address to use (as happens w/ baremetal/ironic in nova) and neutron for some reason still has things assigned with that mac. This is the offending code: 174 except neutron_client_exc.NeutronClientException as e: 175 # NOTE(mriedem): OverQuota in neutron is a 409 176 if e.status_code == 409: 177 LOG.warning(_('Neutron error: quota exceeded')) 178 raise exception.PortLimitExceeded() To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1347778/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1347778] [NEW] raising Maximum number of ports exceeded is wrong
Public bug reported: When neutron API in nova calls create_port(), it looks for exceptions. Any 409 is turned into 'Maximum number of ports exceeded'. This is a horrible assumption. Neutron can return 409s for more than just this reason. Another case where neutron returns a 409 is this: 2014-07-22 18:10:27.583 26577 INFO neutron.api.v2.resource [req-b7267ae5-bafa-4c34-8e25-9c0fca96ad2d None] create failed (client error): Unable to complete operation for network ----. The mac address XX:XX:XX:XX:XX:XX is in use. This can occur when the request to create a port includes the mac address to use (as happens w/ baremetal/ironic in nova) and neutron for some reason still has things assigned with that mac. This is the offending code: 174 except neutron_client_exc.NeutronClientException as e: 175 # NOTE(mriedem): OverQuota in neutron is a 409 176 if e.status_code == 409: 177 LOG.warning(_('Neutron error: quota exceeded')) 178 raise exception.PortLimitExceeded() ** Affects: nova Importance: Undecided Status: New -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1347778 Title: raising Maximum number of ports exceeded is wrong Status in OpenStack Compute (Nova): New Bug description: When neutron API in nova calls create_port(), it looks for exceptions. Any 409 is turned into 'Maximum number of ports exceeded'. This is a horrible assumption. Neutron can return 409s for more than just this reason. Another case where neutron returns a 409 is this: 2014-07-22 18:10:27.583 26577 INFO neutron.api.v2.resource [req-b7267ae5-bafa-4c34-8e25-9c0fca96ad2d None] create failed (client error): Unable to complete operation for network ----. The mac address XX:XX:XX:XX:XX:XX is in use. This can occur when the request to create a port includes the mac address to use (as happens w/ baremetal/ironic in nova) and neutron for some reason still has things assigned with that mac. This is the offending code: 174 except neutron_client_exc.NeutronClientException as e: 175 # NOTE(mriedem): OverQuota in neutron is a 409 176 if e.status_code == 409: 177 LOG.warning(_('Neutron error: quota exceeded')) 178 raise exception.PortLimitExceeded() To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1347778/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1336377] [NEW] random nova unit test failure for test_shelve_offload
Public bug reported: http://logs.openstack.org/02/94402/18/check/gate-nova-python26/b20aa1d/testr_results.html.gz ** Affects: nova Importance: Undecided Status: New -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1336377 Title: random nova unit test failure for test_shelve_offload Status in OpenStack Compute (Nova): New Bug description: http://logs.openstack.org/02/94402/18/check/gate-nova-python26/b20aa1d/testr_results.html.gz To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1336377/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1200591] Re: virt xenapi driver does not retry upload_image on a socket error
The retry code does not check for this: 2014-04-19 00:37:39.354 13204 ERROR nova.compute.manager [req-e7e92354-6e42-4955-9519-08a65872372d ] [instance: 7a2b7c97-f793-4666-888d-430dXXX] Error: [Errno 104] Connection reset by peer in xenapi/client/session.py: def _is_retryable_exception(self, exc, fn): _type, method, error = exc.details[:3] if error == 'RetryableError': LOG.debug(_("RetryableError, so retrying %(fn)s"), {'fn': fn}, exc_info=True) return True elif "signal" in method: LOG.debug(_("Error due to a signal, retrying %(fn)s"), {'fn': fn}, exc_info=True) return True else: return False ** Changed in: nova Status: Invalid => Triaged ** Changed in: nova Assignee: Sridevi Koushik (sridevik) => (unassigned) ** Changed in: nova Importance: Low => Medium -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1200591 Title: virt xenapi driver does not retry upload_image on a socket error Status in OpenStack Compute (Nova): Triaged Bug description: As of now only XenAPI failures are retried in the upload_image method (for the glance_num_retries ). We need to retry on "error: [Errno 104] Connection reset by peer". To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1200591/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1296414] [NEW] quotas not updated when periodic tasks or startup finish deletes
Public bug reported: There are a couple of cases in the compute manager where we don't pass reservations to _delete_instance(). For example, one of them is cleaning up when we see a delete that is stuck in DELETING. The only place we ever update quotas as part of delete should be when the instance DB record is removed. If something is stuck in DELETING, it means that the quota was not updated. We should make sure we're always updating the quota when the instance DB record is removed. (Soft delete kinda throws a wrench in this, hrmph, because I think you want soft deleted instances to not count against quotas -- yet their DB records will still exist.) ** Affects: nova Importance: Undecided Status: New ** Description changed: There are a couple of cases in the compute manager where we don't pass reservations to _delete_instance(). For example, one of them is cleaning up when we see a delete that is stuck in DELETING. The only place we ever update quotas as part of delete should be when the instance DB record is removed. If something is stuck in DELETING, it means that the quota was not updated. We should make sure we're always updating the quota when the instance DB record is removed. + + (Soft delete kinda throws a wrench in this, hrmph, because I think you + want soft deleted instances to not count against quotas -- yet their DB + records will still exist.) -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1296414 Title: quotas not updated when periodic tasks or startup finish deletes Status in OpenStack Compute (Nova): New Bug description: There are a couple of cases in the compute manager where we don't pass reservations to _delete_instance(). For example, one of them is cleaning up when we see a delete that is stuck in DELETING. The only place we ever update quotas as part of delete should be when the instance DB record is removed. If something is stuck in DELETING, it means that the quota was not updated. We should make sure we're always updating the quota when the instance DB record is removed. (Soft delete kinda throws a wrench in this, hrmph, because I think you want soft deleted instances to not count against quotas -- yet their DB records will still exist.) To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1296414/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1212028] Re: SchedulerManager._expire_reservations doesn't need to always be running
This bug is not valid. The task needs to run periodically to expire any reservations that should be expired ahead of QUOTAS.reserve() calls. The only alternative is to expire reservations explicitly before every single QUOTAS.reserve()... which is not as performant. This task doesn't necessarily belong in nova-scheduler. It really could easily be moved to nova-conductor or nova-compute or something... but it needs to run somewhere! ** Changed in: nova Status: Confirmed => Invalid -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1212028 Title: SchedulerManager._expire_reservations doesn't need to always be running Status in OpenStack Compute (Nova): Invalid Bug description: The nova-scheduler log is filled with: 2013-08-13 23:30:56.219 730 DEBUG nova.openstack.common.loopingcall [-] Dynamic looping call sleeping for 60.00 seconds _inner /opt/stack/nova/nova/openstack/common/loopingcall.py:132 2013-08-13 23:31:56.228 730 DEBUG nova.openstack.common.periodic_task [-] Running periodic task SchedulerManager._expire_reservations run_periodic_tasks /opt/stack/nova/nova/openstack/common/periodic_task.py:176 This periodic task is to help the quota system work, but if the scheduler hasn't done anything for the past few minutes there is no reason to keep running this task. Instead it can be restarted upon the next scheduling event. The main benefit from fixing this will be cleaner nova-scheduler log files To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1212028/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1211702] Re: Server list filtered by tenant_id doesn't work
If you're an admin, you must be explicit in saying "I want to reach outside of my project id" by using all_tenants=1. ** Changed in: nova Status: In Progress => Invalid -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1211702 Title: Server list filtered by tenant_id doesn't work Status in OpenStack Compute (Nova): Invalid Bug description: When an admin user try to get the list of servers filtered by tenant_id the filter is overwritten by the admin tenant_id so the actual filter specified is not used but the list is just reporting the the admin user's servers. The problem is in that in the servers API the "project_id' is defined to be equal to the context.project_id: if 'all_tenants' not in search_opts: if context.project_id: search_opts['project_id'] = context.project_id Then in the nova/compute/api.py we have a filter_mapping to associate the tenant_id filter as project_id filter: # search_option to filter_name mapping. filter_mapping = { 'image': 'image_ref', 'name': 'display_name', 'tenant_id': 'project_id', 'flavor': _remap_flavor_filter, 'fixed_ip': _remap_fixed_ip_filter} because of the mapping the filter tenant_id is mapped as project_id then the value project_id in the dictionary is overwritten by the project_id filter defined at the API layer. My solution is basically avoid to add a filter if the same filter has been required in the original request. To do that the idea is, before adding a value in the serach_opts, we need to check if the key we are going to add is present as value in the filter_mapping dict, if so we can force the filter only if the user is not requiring the same value in the original call. In my opinion we should also to move all the logic about the creation and management of the search_opts dictionary in the API code in order to have a single point where the dictionary is created and the let the compute api code just to give results back based on the filters passed. Steps to reproduce: - create an instance (inst1) for the tenant_id X - with an admin user call the servers/detail?tenant_id=X - verify that the instance inst1 is not returned. (Note that if the admin user has some instances the call will get them back) To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1211702/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1179395] Re: Powervm driver does not support IBM i
** Changed in: nova Importance: Undecided => Wishlist ** Changed in: nova Status: Invalid => Confirmed -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1179395 Title: Powervm driver does not support IBM i Status in OpenStack Compute (Nova): Confirmed Bug description: Rationale: Currently OpenStack does not support IBM i in powervm driver. Although the underground virtualization mechanism of IBM i is the same as current supported PowerLinux in PowerVM, IBM i instances can not be deployed correctly. Overview: To support IBM i in powervm, the implementation of the enhancement will be based on current powervm implementation, which includes IVM managed system, single VIOS support and logical volume support. It enrolls a difference in image meta to distinguish IBM i image and Power Linux/AIX. PowerVM will distinguish OS type based on image meta to generate partition accordingly. Use Cases: 1. Create new IBM i instance based on file-based IBM i image. 2. Start IBM i instance. 3. Stop IBM i instance. 4. Create snapshot on a IBM i instance. 5. Delete an IBM i instance. Design: The changes for IBM i supporting are within powervm of Nova, which may include: 1. Specify partition property lpar_env for IBM i instance. 2. Specify partition property ipl_source as b for IBM i instance. 3. Specify LPA_PROFILE_ATTRIBUTES load_source_slot and console_slot for IBM i instance. 4. Specify partition keylock position as norm during powering on IBM i instance. 5. Check lpar_started status for IBM i instances according to refcode and running state. To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1179395/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1186401] Re: instance dict is being abused to pass config_drive_id to libvirt on build
sigh. click too fast on status and you get undesired results. heh. ** Changed in: nova Assignee: (unassigned) => Chris Behrens (cbehrens) ** Changed in: nova Status: Triaged => Won't Fix ** Changed in: nova Status: Won't Fix => In Progress -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1186401 Title: instance dict is being abused to pass config_drive_id to libvirt on build Status in OpenStack Compute (Nova): In Progress Bug description: compute/api create() shoves 'config_drive_id' into instance properties... which get passed down during a build.. but that's not a column in the DB. It looks like it actually makes it all the way into libvirt because we don't use the results from any instance_update() in between, which is bad because the instance we pass has old task_state, etc. Also: This will not work when we switch to passing objects. To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1186401/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1153827] Re: soft delete values in InstanceSystemMetadata on instance delete
Re-opening this bug as it is not actually fixed in h-1. The previous fix needed to be reverted due to bug 1185190. ** Changed in: nova Status: Fix Released => Triaged ** Changed in: nova Assignee: Joe Gordon (jogo) => (unassigned) ** Changed in: nova Milestone: havana-1 => None -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1153827 Title: soft delete values in InstanceSystemMetadata on instance delete Status in OpenStack Compute (Nova): Triaged Bug description: soft delete values in InstanceSystemMetadata on instance delete. Currently InstanceSystemMetadata is used in notify_usage_exists_deleted_instance. Non-deleted systemInstanceMetadta is currently used (with deleted instances). https://github.com/openstack/nova/blob/master/nova/db/sqlalchemy/models.py#L755 In order to fix this the read_deleted flag needs to work for joined tables. To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1153827/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1181412] Re: nova unit tests pass with LOG.*() calls that should generate KeyError
Sigh. This appears to be how the python logging module works. http://docs.python.org/2/library/logging.html Under logger.debug they have an example use of 'extra'...and then under it states: "If you choose to use these attributes in logged messages, you need to exercise some care. In the above example, for instance, the Formatter has been set up with a format string which expects ‘clientip’ and ‘user’ in the attribute dictionary of the LogRecord. If these are missing, the message will not be logged because a string formatting exception will occur. So in this case, you always need to pass the extra dictionary with these keys." :( ** Changed in: nova Status: New => Invalid ** Changed in: oslo Status: New => Invalid -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1181412 Title: nova unit tests pass with LOG.*() calls that should generate KeyError Status in OpenStack Compute (Nova): Invalid Status in Oslo - a Library of Common OpenStack Code: Invalid Bug description: I just found a broken LOG.debug() message in the cells/scheduler which references a non-existant keyword arg. Tests run fine. I've inserted a different completely bogus LOG.debug() message and it also passes. It seems that LOG.debug()s are being turned into no-ops. To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1181412/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1157330] Re: error: 'is not a valid node managed by this compute host.' under load
Not sure if this is truly a bug or not. I can't seem to reproduce it on a fresh compute_nodes table. However, if I have a compute_nodes entry leftover... from switching from XenAPI -> fake... I see this problem just trying to build *1* instance. The problem is that the 'nodename' changes. The fake driver advertises a 'fake-mini' nodename. This may help if you switch from kvm -> fake: http://paste.openstack.org/show/34095/ That should make the fake driver use the same nodename as was probably used for kvm. But ultimately, I'm not sure just switching the driver on a host should "just work". XenAPI's nodename is the host in dom0. kvm's nodename is the same as the nova-compute hostname. There's all sorts of things that may need to be cleaned when swapping the driver. I would try repeating this when an empty compute nodes table and see if this is still a bug. ** Changed in: nova Status: Confirmed => Won't Fix ** Changed in: nova Status: Won't Fix => Triaged -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1157330 Title: error: 'is not a valid node managed by this compute host.' under load Status in OpenStack Compute (Nova): Triaged Bug description: Using Devstack (Grizzly-3) and nova SHA1 5aec4001eb30197f88467e0519203a52c3acd431. And setting 'compute_driver = nova.virt.fake.FakeDriver' in nova.conf vagrant@precise64:~$ nova image-list +--+-+++ | ID | Name| Status | Server | +--+-+++ | 90e59cc9-503f-46e8-873e-13cde01b1252 | cirros-0.3.1-x86_64-uec | ACTIVE || | 8097952f-6ef8-440b-bdf4-dfac34746396 | cirros-0.3.1-x86_64-uec-kernel | ACTIVE || | 23828452-dc20-4e4e-a505-86abc77a4cc3 | cirros-0.3.1-x86_64-uec-ramdisk | ACTIVE || +--+-+++ When trying to run 100 vms as admin user using $ nova boot --image 90e59cc9-503f-46e8-873e-13cde01b1252 --flavor m1.nano --num-instances 100 test over %10 fail with the following error: {u'message': u'NovaException', u'code': 500, u'details': u'precise64 is not a valid node managed by this compute host. http://paste.openstack.org/show/34069/ To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1157330/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1142901] Re: keypairs aren't propagated with cells
Ah, key data is sent with each instance already... so we really don't need this in child cells. ** Changed in: nova Status: In Progress => Invalid -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1142901 Title: keypairs aren't propagated with cells Status in OpenStack Compute (Nova): Invalid Bug description: Keypairs don't work with cells as they are not synced down to child cells. To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1142901/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Bug 1130997] Re: host_power_action doesn't work in cells
Turns out this is not a bug. There's a bit of trickery in cells_api's version of HostAPI() in that unimplemented methods there use api.py's HostAPI() versions of methods and the rpcapi class is swapped out for one that proxies via cells to the correct cell and manager. ** Changed in: nova Status: In Progress => Invalid -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to OpenStack Compute (nova). https://bugs.launchpad.net/bugs/1130997 Title: host_power_action doesn't work in cells Status in OpenStack Compute (Nova): Invalid Bug description: nova.compute.api.HostAPI.host_power_action needs to be added to nova.compute.cells_api At the moment if you have a host in a different cell to your api server, you can't call host_power_action on it. To manage notifications about this bug go to: https://bugs.launchpad.net/nova/+bug/1130997/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Blueprint scaling-zones] Improvements for Scaling Zones
Blueprint changed by Chris Behrens: Assignee: (none) => Nova Scaling Team -- Improvements for Scaling Zones https://blueprints.launchpad.net/nova/+spec/scaling-zones -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp
[Yahoo-eng-team] [Blueprint scaling-zones] Improvements for Scaling Zones
Blueprint changed by Chris Behrens: Drafter: Chris Behrens => Nova Scaling Team -- Improvements for Scaling Zones https://blueprints.launchpad.net/nova/+spec/scaling-zones -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp