Re: [Openstack] [Ceilometer] Problem with nova resize events
Hi Jay and Julien thanks very much for your help. I am not terribly familiar with github and code reviews, I am not a developer :(, so maybe I am not using the correct terms here. We are using version 2013.1.2 of ceilometer, so we have no ComputeInstanceNotificationBase class in notifications.py, instead we have class _Base(plugin.NotificationBase) , which handles events with get_event_types(): return ['compute.instance.create.end', 'compute.instance.exists', 'compute.instance.delete.start', 'compute.instance.finish_resize.end', 'compute.instance.resize.revert.end'] like ComputeInstanceNotificationBase does. What is your suggestion in order to enable handlings of more compute events, considering we will go live in 2-3 weeks time? 1- We wait for Julien's patch to be approved and we upgrade to the patched version. 2- We modify ourselves get_event_types() on our installation 3- ? Thanks Alex -Original Message- From: Julien Danjou [mailto:jul...@danjou.info] Sent: mercoledì 24 luglio 2013 17:39 To: Jay Lau Cc: Alessandro Barabesi; openstack@lists.launchpad.net Subject: Re: [Openstack] [Ceilometer] Problem with nova resize events On Wed, Jul 24 2013, Jay Lau wrote: Hi Alex, Not sure if it is a bug, but you are right, ceiloemeter did not send the notification that you required. Perhaps you can log a bug or else you can patch your cluster directly to enable this. class ComputeInstanceNotificationBase(ComputeNotificationBase): Convert compute.instance.* notifications into Counters @staticmethod def get_event_types(): return ['compute.instance.create.start', 'compute.instance.create.end', 'compute.instance.exists', 'compute.instance.update', 'compute.instance.delete.start', 'compute.instance.delete.end', 'compute.instance.finish_resize.end', 'compute.instance.resize.revert.end'] Thanks, Not really a bug, but we know we could do better. I've just cooked a patch that should improve that: https://review.openstack.org/38485 -- Julien Danjou # Free Software hacker # freelance consultant # http://julien.danjou.info ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
[Openstack] [Ceilometer] Problems with query fields in filters
Hi everybody, I have the following problems with query fields in filters. 1) Filtering by metadata.size in v2/meters/image apparently is not working. The following request returns also samples with metadata.size=0. GET http://10.10.10.10:8777/v2/meters/image { q: [{ field: project_id, op: eq, value: 77b461539c8542909f67b29939ec87dd }, { field: timestamp, op: ge, value: 2013-07-11T13:36:00 }, { field: timestamp, op: lt, value: 2013-07-11T13:39:00 }, { field: metadata.size, op: gt, value: 0 }] } I am probably doing something wrong but I can't figure out what. 2) Filtering by counter_volume returns the folloving error message: error_message={debuginfo: null, faultcode: Client, faultstring: Unknown argument: \counter_volume\: unrecognized query field} Is this a bug or filtering by counter-volume is not allowed? Many thanks Alex ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] [Ceilometer] Problems with query fields in filters
Hi Julien thanks for your reply, unfortunately volume instead of counter_volume is not working either. Alex -Original Message- From: Julien Danjou [mailto:jul...@danjou.info] Sent: venerdì 12 luglio 2013 11:34 To: Alessandro Barabesi Cc: openstack@lists.launchpad.net Subject: Re: [Openstack] [Ceilometer] Problems with query fields in filters On Fri, Jul 12 2013, Alessandro Barabesi wrote: I have the following problems with query fields in filters. 1) Filtering by metadata.size in v2/meters/image apparently is not working. The following request returns also samples with metadata.size=0. GET http://10.10.10.10:8777/v2/meters/image { q: [{ field: project_id, op: eq, value: 77b461539c8542909f67b29939ec87dd }, { field: timestamp, op: ge, value: 2013-07-11T13:36:00 }, { field: timestamp, op: lt, value: 2013-07-11T13:39:00 }, { field: metadata.size, op: gt, value: 0 }] } I am probably doing something wrong but I can't figure out what. That seems correct from the top of my head. If that really doesn't work, feel free to open a bug. 2) Filtering by counter_volume returns the folloving error message: error_message={debuginfo: null, faultcode: Client, faultstring: Unknown argument: \counter_volume\: unrecognized query field} Is this a bug or filtering by counter-volume is not allowed? Try `volume' instead of `counter_volume'. But I am not sure it's currently allowed -- in that case opening a bug can be good idea too. -- Julien Danjou // Free Software hacker / freelance consultant // http://julien.danjou.info ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
[Openstack] [Ceilometer] volume.exists events collected after volume is deleted
Hi everybody, ceilometer collects samples for a volume even after the volume has been deleted, reporting event volume.exists (see below). Does it dependend on ceilometer or on cinder? Is there a way to stop this? [{ counter_name: volume.size, user_id: 0ac98027a16849339a31f4f1f1114df2, resource_id: 44e2cf92-5f72-49c6-b0d7-4f8ec2c0bde9, timestamp: 2013-07-11T16:24:02.239000, resource_metadata: { status: deleted, display_name: Vol_02_Ale_Test_03, event_type: volume.exists, volume_type: None, host: volume.grizzly, size: 6 }, source: openstack, counter_unit: B, counter_volume: 6.0, project_id: 77b461539c8542909f67b29939ec87dd, message_id: 4c40cffc-ea46-11e2-9a49-00259075703e, counter_type: gauge }, { counter_name: volume.size, user_id: 0ac98027a16849339a31f4f1f1114df2, resource_id: 6267f31c-293a-43ec-b691-b81742364c92, timestamp: 2013-07-11T16:24:02.249000, resource_metadata: { status: in-use, display_name: Vol_01_Ale_Test_03, event_type: volume.exists, volume_type: None, host: volume.grizzly, size: 3 }, source: openstack, counter_unit: B, counter_volume: 3.0, project_id: 77b461539c8542909f67b29939ec87dd, message_id: 4c425610-ea46-11e2-9a49-00259075703e, counter_type: gauge }] Thanks Alex ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] [Ceilometer] Problems with query fields in filters
Hi Angus without the metadata.size filter I get all samples, with sizeo and with size=0. Same thing with the metadata.size filter using the following operators: gt,lt,ne,le,ge The most strange thing is that if I use the operator eq I get an empty response! Thanks Alex -Original Message- From: Angus Salkeld [mailto:asalk...@redhat.com] Sent: venerdì 12 luglio 2013 12:55 To: openstack@lists.launchpad.net Cc: Alessandro Barabesi Subject: Re: [Openstack] [Ceilometer] Problems with query fields in filters On 12/07/13 11:33 +0200, Julien Danjou wrote: On Fri, Jul 12 2013, Alessandro Barabesi wrote: I have the following problems with query fields in filters. 1) Filtering by metadata.size in v2/meters/image apparently is not working. The following request returns also samples with metadata.size=0. GET http://10.10.10.10:8777/v2/meters/image { q: [{ field: project_id, op: eq, value: 77b461539c8542909f67b29939ec87dd }, { field: timestamp, op: ge, value: 2013-07-11T13:36:00 }, { field: timestamp, op: lt, value: 2013-07-11T13:39:00 }, { field: metadata.size, op: gt, value: 0 I suspect the gt operator is not working (it's probably using eq given what you are getting). But certainly a bug. I'd just remove this last query and see what you get with the first 3 queries. -Angus }] } I am probably doing something wrong but I can't figure out what. That seems correct from the top of my head. If that really doesn't work, feel free to open a bug. 2) Filtering by counter_volume returns the folloving error message: error_message={debuginfo: null, faultcode: Client, faultstring: Unknown argument: \counter_volume\: unrecognized query field} Is this a bug or filtering by counter-volume is not allowed? Try `volume' instead of `counter_volume'. But I am not sure it's currently allowed -- in that case opening a bug can be good idea too. -- Julien Danjou // Free Software hacker / freelance consultant // http://julien.danjou.info ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp