1 Detailed architecture and interface specification
2 =================================================
4 This section describes a detailed implementation plan, which is based on the
5 high level architecture introduced in Section 3. Section 5.1 describes the
6 functional blocks of the Doctor architecture, which is followed by a high level
7 message flow in Section 5.2. Section 5.3 provides a mapping of selected existing
8 open source components to the building blocks of the Doctor architecture.
9 Thereby, the selection of components is based on their maturity and the gap
10 analysis executed in Section 4. Sections 5.4 and 5.5 detail the specification of
11 the related northbound interface and the related information elements. Finally,
12 Section 5.6 provides a first set of blueprints to address selected gaps required
13 for the realization functionalities of the Doctor project.
20 This section introduces the functional blocks to form the VIM. OpenStack was
21 selected as the candidate for implementation. Inside the VIM, 4 different
22 building blocks are defined (see :num:`Figure #figure6`).
26 .. figure:: images/figure6.png
34 The Monitor module has the responsibility for monitoring the virtualized
35 infrastructure. There are already many existing tools and services (e.g. Zabbix)
36 to monitor different aspects of hardware and software resources which can be
37 used for this purpose.
42 The Inspector module has the ability a) to receive various failure notifications
43 regarding physical resource(s) from Monitor module(s), b) to find the affected
44 virtual resource(s) by querying the resource map in the Controller, and c) to
45 update the state of the virtual resource (and physical resource).
47 The Inspector has drivers for different types of events and resources to
48 integrate any type of Monitor and Controller modules. It also uses a failure
49 policy database to decide on the failure selection and aggregation from raw
50 events. This failure policy database is configured by the Administrator.
52 The reason for separation of the Inspector and Controller modules is to make the
53 Controller focus on simple operations by avoiding a tight integration of various
54 health check mechanisms into the Controller.
59 The Controller is responsible for maintaining the resource map (i.e. the mapping
60 from physical resources to virtual resources), accepting update requests for the
61 resource state(s) (exposing as provider API), and sending all failure events
62 regarding virtual resources to the Notifier. Optionally, the Controller has the
63 ability to force the state of a given physical resource to down in the resource
64 mapping when it receives failure notifications from the Inspector for that
65 given physical resource.
66 The Controller also re-calculates the capacity of the NVFI when receiving a
67 failure notification for a physical resource.
69 In a real-world deployment, the VIM may have several controllers, one for each
70 resource type, such as Nova, Neutron and Cinder in OpenStack. Each controller
71 maintains a database of virtual and physical resources which shall be the master
72 source for resource information inside the VIM.
77 The focus of the Notifier is on selecting and aggregating failure events
78 received from the controller based on policies mandated by the Consumer.
79 Therefore, it allows the Consumer to subscribe for alarms regarding virtual
80 resources using a method such as API endpoint. After receiving a fault
81 event from a Controller, it will notify the fault to the Consumer by referring
82 to the alarm configuration which was defined by the Consumer earlier on.
84 To reduce complexity of the Controller, it is a good approach for the
85 Controllers to emit all notifications without any filtering mechanism and have
86 another service (i.e. Notifier) handle those notifications properly. This is the
87 general philosophy of notifications in OpenStack. Note that a fault message
88 consumed by the Notifier is different from the fault message received by the
89 Inspector; the former message is related to virtual resources which are visible
90 to users with relevant ownership, whereas the latter is related to raw devices
91 or small entities which should be handled with an administrator privilege.
93 The northbound interface between the Notifier and the Consumer/Administrator is
94 specified in :ref:`impl_nbi`.
102 The detailed work flow for fault management is as follows (see also :num:`Figure
105 1. Request to subscribe to monitor specific virtual resources. A query filter
106 can be used to narrow down the alarms the Consumer wants to be informed
108 2. Each subscription request is acknowledged with a subscribe response message.
109 The response message contains information about the subscribed virtual
110 resources, in particular if a subscribed virtual resource is in "alarm"
112 3. The NFVI sends monitoring events for resources the VIM has been subscribed
113 to. Note: this subscription message exchange between the VIM and NFVI is not
114 shown in this message flow.
115 4. Event correlation, fault detection and aggregation in VIM.
116 5. Database lookup to find the virtual resources affected by the detected fault.
117 6. Fault notification to Consumer.
118 7. The Consumer switches to standby configuration (STBY)
119 8. Instructions to VIM requesting certain actions to be performed on the
120 affected resources, for example migrate/update/terminate specific
121 resource(s). After reception of such instructions, the VIM is executing the
122 requested action, e.g. it will migrate or terminate a virtual resource.
124 a. Query request from Consumer to VIM to get information about the current
125 status of a resource.
126 b. Response to the query request with information about the current status of
127 the queried resource. In case the resource is in "fault" state, information
128 about the related fault(s) is returned.
130 In order to allow for quick reaction to failures, the time interval between
131 fault detection in step 3 and the corresponding recovery actions in step 7 and 8
132 shall be less than 1 second.
136 .. figure:: images/figure7.png
139 Fault management work flow
144 .. figure:: images/figure8.png
147 Fault management scenario
149 :num:`Figure #figure8` shows a more detailed message flow (Steps 4 to 6) between
150 the 4 building blocks introduced in :ref:`impl_fb`.
152 4. The Monitor observed a fault in the NFVI and reports the raw fault to the
154 The Inspector filters and aggregates the faults using pre-configured
158 a) The Inspector queries the Resource Map to find the virtual resources
159 affected by the raw fault in the NFVI.
160 b) The Inspector updates the state of the affected virtual resources in the
162 c) The Controller observes a change of the virtual resource state and informs
163 the Notifier about the state change and the related alarm(s).
164 Alternatively, the Inspector may directly inform the Notifier about it.
166 6. The Notifier is performing another filtering and aggregation of the changes
167 and alarms based on the pre-configured alarm configuration. Finally, a fault
168 notification is sent to northbound to the Consumer.
173 The detailed work flow for NFVI maintenance is shown in :num:`Figure #figure9`
174 and has the following steps. Note that steps 1, 2, and 5 to 8a in the NFVI
175 maintenance work flow are very similar to the steps in the fault management work
176 flow and share a similar implementation plan in Release 1.
178 1. Subscribe to fault/maintenance notifications.
179 2. Response to subscribe request.
180 3. Maintenance trigger received from administrator.
181 4. VIM switches NFVI resources to "maintenance" state. This, e.g., means they
182 should not be used for further allocation/migration requests
183 5. Database lookup to find the virtual resources affected by the detected
184 maintenance operation.
185 6. Maintenance notification to Consumer.
186 7. The Consumer switches to standby configuration (STBY)
187 8. Instructions from Consumer to VIM requesting certain recovery actions to be
188 performed (step 7a). After reception of such instructions, the VIM is
189 executing the requested action in order to empty the physical resources (step
191 9. Maintenance response from VIM to inform the Administrator that the physical
192 machines have been emptied (or the operation resulted in an error state).
193 10. Administrator is coordinating and executing the maintenance operation/work
196 A) Query request from Administrator to VIM to get information about the
197 current state of a resource.
198 B) Response to the query request with information about the current state of
199 the queried resource(s). In case the resource is in "maintenance" state,
200 information about the related maintenance operation is returned.
204 .. figure:: images/figure9.png
207 NFVI maintenance work flow
212 .. figure:: images/figure10.png
215 NFVI Maintenance implementation plan
217 :num:`Figure #figure10` shows a more detailed message flow (Steps 4 to 6)
218 between the 4 building blocks introduced in Section 5.1..
220 3. The Administrator is sending a StateChange request to the Controller residing
222 4. The Controller queries the Resource Map to find the virtual resources
223 affected by the planned maintenance operation.
226 a) The Controller updates the state of the affected virtual resources in the
227 Resource Map database.
229 b) The Controller informs the Notifier about the virtual resources that will
230 be affected by the maintenance operation.
232 6. A maintenance notification is sent to northbound to the Consumer.
236 9. The Controller informs the Administrator after the physical resources have
241 Implementation plan for OPNFV Release 1
242 ---------------------------------------
247 :num:`Figure #figure11` shows the implementation plan based on OpenStack and
248 related components as planned for Release 1. Hereby, the Monitor can be realized
249 by Zabbix. The Controller is realized by OpenStack Nova [NOVA]_, Neutron
250 [NEUT]_, and Cinder [CIND]_ for compute, network, and storage,
251 respectively. The Inspector can be realized by Monasca [MONA]_ or a simple
252 script querying Nova in order to map between physical and virtual resources. The
253 Notifier will be realized by Ceilometer [CEIL]_ receiving failure events
254 on its notification bus.
256 :num:`Figure #figure12` shows the inner-workings of Ceilometer. After receiving
257 an "event" on its notification bus, first a notification agent will grab the
258 event and send a "notification" to the Collector. The collector writes the
259 notifications received to the Ceilometer databases.
261 In the existing Ceilometer implementation, an alarm evaluator is periodically
262 polling those databases through the APIs provided. If it finds new alarms, it
263 will evaluate them based on the pre-defined alarm configuration, and depending
264 on the configuration, it will hand a message to the Alarm Notifier, which in
265 turn will send the alarm message northbound to the Consumer. :num:`Figure
266 #figure12` also shows an optimized work flow for Ceilometer with the goal to
267 reduce the delay for fault notifications to the Consumer. The approach is to
268 implement a new notification agent (called "publisher" in Ceilometer
269 terminology) which is directly sending the alarm through the "Notification Bus"
270 to a new "Notification-driven Alarm Evaluator (NAE)" (see Sections 5.6.2 and
271 5.6.3), thereby bypassing the Collector and avoiding the additional delay of the
272 existing polling-based alarm evaluator. The NAE is similar to the OpenStack
273 "Alarm Evaluator", but is triggered by incoming notifications instead of
274 periodically polling the OpenStack "Alarms" database for new alarms. The
275 Ceilometer "Alarms" database can hold three states: "normal", "insufficient
276 data", and "fired". It is representing a persistent alarm database. In order to
277 realize the Doctor requirements, we need to define new "meters" in the database
282 .. figure:: images/figure11.png
285 Implementation plan in OpenStack (OPNFV Release 1 ”Arno”)
290 .. figure:: images/figure12.png
293 Implementation plan in Ceilometer architecture
299 For NFVI Maintenance, a quite similar implementation plan exists. Instead of a
300 raw fault being observed by the Monitor, the Administrator is sending a
301 Maintenance Request through the northbound interface towards the Controller
302 residing in the VIM. Similar to the Fault Management use case, the Controller
303 (in our case OpenStack Nova) will send a maintenance event to the Notifier (i.e.
304 Ceilometer in our implementation). Within Ceilometer, the same workflow as
305 described in the previous section applies. In addition, the Controller(s) will
306 take appropriate actions to evacuate the physical machines in order to prepare
307 them for the planned maintenance operation. After the physical machines are
308 emptied, the Controller will inform the Administrator that it can initiate the
314 This section introduces all attributes and information elements used in the
315 messages exchange on the northbound interfaces between the VIM and the VNFO and
318 Note: The information elements will be aligned with current work in ETSI NFV IFA
322 Simple information elements:
324 * SubscriptionID: identifies a subscription to receive fault or maintenance
326 * NotificationID: identifies a fault or maintenance notification.
327 * VirtualResourceID (Identifier): identifies a virtual resource affected by a
328 fault or a maintenance action of the underlying physical resource.
329 * PhysicalResourceID (Identifier): identifies a physical resource affected by a
330 fault or maintenance action.
331 * VirtualResourceState (String): state of a virtual resource, e.g. "normal",
332 "maintenance", "down", "error".
333 * PhysicalResourceState (String): state of a physical resource, e.g. "normal",
334 "maintenance", "down", "error".
335 * VirtualResourceType (String): type of the virtual resource, e.g. "virtual
336 machine", "virtual memory", "virtual storage", "virtual CPU", or "virtual
338 * FaultID (Identifier): identifies the related fault in the underlying physical
339 resource. This can be used to correlate different fault notifications caused
340 by the same fault in the physical resource.
341 * FaultType (String): Type of the fault. The allowed values for this parameter
342 depend on the type of the related physical resource. For example, a resource
343 of type "compute hardware" may have faults of type "CPU failure", "memory
344 failure", "network card failure", etc.
345 * Severity (Integer): value expressing the severity of the fault. The higher the
346 value, the more severe the fault.
347 * MinSeverity (Integer): value used in filter information elements. Only faults
348 with a severity higher than the MinSeverity value will be notified to the
350 * EventTime (Datetime): Time when the fault was observed.
351 * EventStartTime and EventEndTime (Datetime): Datetime range that can be used in
352 a FaultQueryFilter to narrow down the faults to be queried.
353 * ProbableCause: information about the probable cause of the fault.
354 * CorrelatedFaultID (Integer): list of other faults correlated to this fault.
355 * isRootCause (Boolean): Parameter indicating if this fault is the root for
356 other correlated faults. If TRUE, then the faults listed in the parameter
357 CorrelatedFaultID are caused by this fault.
358 * FaultDetails (Key-value pair): provides additional information about the
359 fault, e.g. information about the threshold, monitored attributes, indication
360 of the trend of the monitored parameter.
361 * FirmwareVersion (String): current version of the firmware of a physical
363 * HypervisorVersion (String): current version of a hypervisor.
364 * ZoneID (Identifier): Identifier of the resource zone. A resource zone is the
365 logical separation of physical and software resources in an NFVI deployment
366 for physical isolation, redundancy, or administrative designation.
367 * Metadata (Key-Value-Pairs): provides additional information of a physical
368 resource in maintenance/error state.
370 Complex information elements (see also UML diagrams in :num:`Figure #figure13`
371 and :num:`Figure #figure14`):
373 * VirtualResourceInfoClass:
375 + VirtualResourceID [1] (Identifier)
376 + VirtualResourceState [1] (String)
377 + Faults [0..*] (FaultClass): For each resource, all faults
378 including detailed information about the faults are provided.
380 * FaultClass: The parameters of the FaultClass are partially based on ETSI TS
381 132 111-2 (V12.1.0) [*]_, which is specifying fault management in 3GPP, in
382 particular describing the information elements used for alarm notifications.
384 - FaultID [1] (Identifier)
386 - Severity [1] (Integer)
387 - EventTime [1] (Datetime)
389 - CorrelatedFaultID [0..*] (Identifier)
390 - FaultDetails [0..*] (Key-value pair)
392 .. [*] http://www.etsi.org/deliver/etsi_ts/132100_132199/13211102/12.01.00_60/ts_13211102v120100p.pdf
394 * SubscribeFilterClass
396 - VirtualResourceType [0..*] (String)
397 - VirtualResourceID [0..*] (Identifier)
398 - FaultType [0..*] (String)
399 - MinSeverity [0..1] (Integer)
401 * FaultQueryFilterClass: narrows down the FaultQueryRequest, for example it
402 limits the query to certain physical resources, a certain zone, a given fault
403 type/severity/cause, or a specific FaultID.
405 - VirtualResourceType [0..*] (String)
406 - VirtualResourceID [0..*] (Identifier)
407 - FaultType [0..*] (String)
408 - MinSeverity [0..1] (Integer)
409 - EventStartTime [0..1] (Datetime)
410 - EventEndTime [0..1] (Datetime)
412 * PhysicalResourceStateClass:
414 - PhysicalResourceID [1] (Identifier)
415 - PhysicalResourceState [1] (String): mandates the new state of the physical
418 * PhysicalResourceInfoClass:
420 - PhysicalResourceID [1] (Identifier)
421 - PhysicalResourceState [1] (String)
422 - FirmwareVersion [0..1] (String)
423 - HypervisorVersion [0..1] (String)
424 - ZoneID [0..1] (Identifier)
426 * StateQueryFilterClass: narrows down a StateQueryRequest, for example it limits
427 the query to certain physical resources, a certain zone, or a given resource
428 state (e.g., only resources in "maintenance" state).
430 - PhysicalResourceID [1] (Identifier)
431 - PhysicalResourceState [1] (String)
432 - ZoneID [0..1] (Identifier)
436 Detailed northbound interface specification
437 -------------------------------------------
439 This section is specifying the northbound interfaces for fault management and
440 NFVI maintenance between the VIM on the one end and the Consumer and the
441 Administrator on the other ends. For each interface all messages and related
442 information elements are provided.
444 Note: The interface definition will be aligned with current work in ETSI NFV IFA
447 All of the interfaces described below are produced by the VIM and consumed by
448 the Consumer or Administrator.
450 Fault management interface
451 ^^^^^^^^^^^^^^^^^^^^^^^^^^
453 This interface allows the VIM to notify the Consumer about a virtual resource
454 that is affected by a fault, either within the virtual resource itself or by the
455 underlying virtualization infrastructure. The messages on this interface are
456 shown in :num:`Figure #figure13` and explained in detail in the following
459 Note: The information elements used in this section are described in detail in
464 .. figure:: images/figure13.png
467 Fault management NB I/F messages
470 SubscribeRequest (Consumer -> VIM)
471 __________________________________
473 Subscription from Consumer to VIM to be notified about faults of specific
474 resources. The faults to be notified about can be narrowed down using a
479 - SubscribeFilter [1] (SubscribeFilterClass): Optional information to narrow
480 down the faults that shall be notified to the Consumer, for example limit to
481 specific VirtualResourceID(s), severity, or cause of the alarm.
483 SubscribeResponse (VIM -> Consumer)
484 ___________________________________
486 Response to a subscribe request message including information about the
487 subscribed resources, in particular if they are in "fault/error" state.
491 * SubscriptionID [1] (Identifier): Unique identifier for the subscription. It
492 can be used to delete or update the subscription.
493 * VirtualResourceInfo [0..*] (VirtualResourceInfoClass): Provides additional
494 information about the subscribed resources, i.e., a list of the related
495 resources, the current state of the resources, etc.
497 FaultNotification (VIM -> Consumer)
498 ___________________________________
500 Notification about a virtual resource that is affected by a fault, either within
501 the virtual resource itself or by the underlying virtualization infrastructure.
502 After reception of this request, the Consumer will decide on the optimal
503 action to resolve the fault. This includes actions like switching to a hot
504 standby virtual resource, migration of the fault virtual resource to another
505 physical machine, termination of the faulty virtual resource and instantiation
506 of a new virtual resource in order to provide a new hot standby resource.
507 Existing resource management interfaces and messages between the Consumer and
508 the VIM can be used for those actions, and there is no need to define additional
509 actions on the Fault Management Interface.
513 * NotificationID [1] (Identifier): Unique identifier for the notification.
514 * VirtualResourceInfo [1..*] (VirtualResourceInfoClass): List of faulty
515 resources with detailed information about the faults.
517 FaultQueryRequest (Consumer -> VIM)
518 ___________________________________
520 Request to find out about active alarms at the VIM. A FaultQueryFilter can be
521 used to narrow down the alarms returned in the response message.
525 * FaultQueryFilter [1] (FaultQueryFilterClass): narrows down the
526 FaultQueryRequest, for example it limits the query to certain physical
527 resources, a certain zone, a given fault type/severity/cause, or a specific
530 FaultQueryResponse (VIM -> Consumer)
531 ____________________________________
533 List of active alarms at the VIM matching the FaultQueryFilter specified in the
538 * VirtualResourceInfo [0..*] (VirtualResourceInfoClass): List of faulty
539 resources. For each resource all faults including detailed information about
540 the faults are provided.
545 The NFVI maintenance interfaces Consumer-VIM allows the Consumer to subscribe to
546 maintenance notifications provided by the VIM. The related maintenance interface
547 Administrator-VIM allows the Administrator to issue maintenance requests to the
548 VIM, i.e. requesting the VIM to take appropriate actions to empty physical
549 machine(s) in order to execute maintenance operations on them. The interface
550 also allows the Administrator to query the state of physical machines, e.g., in
551 order to get details in the current status of the maintenance operation like a
554 The messages defined in these northbound interfaces are shown in :num:`Figure
555 #figure14` and described in detail in the following subsections.
559 .. figure:: images/figure14.png
562 NFVI maintenance NB I/F messages
564 SubscribeRequest (Consumer -> VIM)
565 __________________________________
567 Subscription from Consumer to VIM to be notified about maintenance operations
568 for specific virtual resources. The resources to be informed about can be
569 narrowed down using a subscribe filter.
573 * SubscribeFilter [1] (SubscribeFilterClass): Information to narrow down the
574 faults that shall be notified to the Consumer, for example limit to specific
575 virtual resource type(s).
577 SubscribeResponse (VIM -> Consumer)
578 ___________________________________
580 Response to a subscribe request message, including information about the
581 subscribed virtual resources, in particular if they are in "maintenance" state.
585 * SubscriptionID [1] (Identifier): Unique identifier for the subscription. It
586 can be used to delete or update the subscription.
587 * VirtualResourceInfo [0..*] (VirtalResourceInfoClass): Provides additional
588 information about the subscribed virtual resource(s), e.g., the ID, type and
589 current state of the resource(s).
591 MaintenanceNotification (VIM -> Consumer)
592 _________________________________________
594 Notification about a physical resource switched to "maintenance" state. After
595 reception of this request, the Consumer will decide on the optimal action to
596 address this request, e.g., to switch to the standby (STBY) configuration.
600 * VirtualResourceInfo [1..*] (VirtualResourceInfoClass): List of virtual
601 resources where the state has been changed to maintenance.
603 StateChangeRequest (Administrator -> VIM)
604 _________________________________________
606 Request to change the state of a list of physical resources, e.g. to
607 "maintenance" state, in order to prepare them for a planned maintenance
612 * PhysicalResourceState [1..*] (PhysicalResourceStateClass)
614 StateChangeResponse (VIM -> Administrator)
615 __________________________________________
617 Response message to inform the Administrator that the requested resources are
618 now in maintenance state (or the operation resulted in an error) and the
619 maintenance operation(s) can be executed.
623 * PhysicalResourceInfo [1..*] (PhysicalResourceInfoClass)
625 StateQueryRequest (Administrator -> VIM)
626 ________________________________________
628 In this procedure, the Administrator would like to get the information about
629 physical machine(s), e.g. their state ("normal", "maintenance"), firmware
630 version, hypervisor version, update status of firmware and hypervisor, etc. It
631 can be used to check the progress during firmware update and the confirmation
632 after update. A filter can be used to narrow down the resources returned in the
637 * StateQueryFilter [1] (StateQueryFilterClass): narrows down the
638 StateQueryRequest, for example it limits the query to certain physical
639 resources, a certain zone, or a given resource state.
641 StateQueryResponse (VIM -> Administrator)
642 _________________________________________
644 List of physical resources matching the filter specified in the
649 * PhysicalResourceInfo [0..*] (PhysicalResourceInfoClass): List of physical
650 resources. For each resource, information about the current state, the
651 firmware version, etc. is provided.
656 This section is listing a first set of blueprints that have been proposed by the
657 Doctor project to the open source community. Further blueprints addressing other
658 gaps identified in Section 4 will be submitted at a later stage of the OPNFV. In
659 this section the following definitions are used:
661 * "Event" is a message emitted by other OpenStack services such as Nova and
662 Neutron and is consumed by the "Notification Agents" in Ceilometer.
663 * "Notification" is a message generated by a "Notification Agent" in Ceilometer
664 based on an "event" and is delivered to the "Collectors" in Ceilometer that
665 store those notifications (as "sample") to the Ceilometer "Databases".
667 Instance State Notification (Ceilometer) [*]_
668 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
670 The Doctor project is planning to handle "events" and "notifications" regarding
671 Resource Status; Instance State, Port State, Host State, etc. Currently,
672 Ceilometer already receives "events" to identify the state of those resources,
673 but it does not handle and store them yet. This is why we also need a new event
674 definition to capture those resource states from "events" created by other
677 This BP proposes to add a new compute notification state to handle events from
678 an instance (server) from nova. It also creates a new meter "instance.state" in
681 .. [*] https://etherpad.opnfv.org/p/doctor_bps
683 Event Publisher for Alarm (Ceilometer) [*]_
684 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
686 **Problem statement:**
688 The existing "Alarm Evaluator" in OpenStack Ceilometer is periodically
689 querying/polling the databases in order to check all alarms independently from
690 other processes. This is adding additional delay to the fault notification
691 send to the Consumer, whereas one requirement of Doctor is to react on faults
694 The existing message flow is shown in :num:`Figure #figure12`: after receiving
695 an "event", a "notification agent" (i.e. "event publisher") will send a
696 "notification" to a "Collector". The "collector" is collecting the
697 notifications and is updating the Ceilometer "Meter" database that is storing
698 information about the "sample" which is capured from original "event". The
699 "Alarm Evaluator" is periodically polling this databases then querying "Meter"
700 database based on each alarm configuration.
702 In the current Ceilometer implementation, there is no possibility to directly
703 trigger the "Alarm Evaluator" when a new "event" was received, but the "Alarm
704 Evaluator" will only find out that requires firing new notification to the
705 Consumer when polling the database.
707 **Change/feature request:**
709 This BP proposes to add a new "event publisher for alarm", which is bypassing
710 several steps in Ceilometer in order to avoid the polling-based approach of
711 the existing Alarm Evaluator that makes notification slow to users.
713 After receiving an "(alarm) event" by listening on the Ceilometer message
714 queue ("notification bus"), the new "event publisher for alarm" immediately
715 hands a "notification" about this event to a new Ceilometer component
716 "Notification-driven alarm evaluator" proposed in the other BP (see Section
719 Note, the term "publisher" refers to an entity in the Ceilometer architecture
720 (it is a "notification agent"). It offers the capability to provide
721 notifications to other services outside of Ceilometer, but it is also used to
722 deliver notifications to other Ceilometer components (e.g. the "Collectors")
723 via the Ceilometer "notification bus".
725 **Implementation detail**
727 * "Event publisher for alarm" is part of Ceilometer
728 * The standard AMQP message queue is used with a new topic string.
729 * No new interfaces have to be added to Ceilometer.
730 * "Event publisher for Alarm" can be configured by the Administrator of
731 Ceilometer to be used as "Notification Agent" in addition to the existing
733 * Existing alarm mechanisms of Ceilometer can be used allowing users to
734 configure how to distribute the "notifications" transformed from "events",
735 e.g. there is an option whether an ongoing alarm is re-issued or not
738 .. [*] https://etherpad.opnfv.org/p/doctor_bps
740 Notification-driven alarm evaluator (Ceilometer) [*]_
741 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
743 **Problem statement:**
745 The existing "Alarm Evaluator" in OpenStack Ceilometer is periodically
746 querying/polling the databases in order to check all alarms independently from
747 other processes. This is adding additional delay to the fault notification send
748 to the Consumer, whereas one requirement of Doctor is to react on faults as fast
751 **Change/feature request:**
753 This BP is proposing to add an alternative "Notification-driven Alarm Evaluator"
754 for Ceilometer that is receiving "notifications" sent by the "Event Publisher
755 for Alarm" described in the other BP. Once this new "Notification-driven Alarm
756 Evaluator" received "notification", it finds the "alarm" configurations which
757 may relate to the "notification" by querying the "alarm" database with some keys
758 i.e. resource ID, then it will evaluate each alarm with the information in that
761 After the alarm evaluation, it will perform the same way as the existing "alarm
762 evaluator" does for firing alarm notification to the Consumer. Similar to the
763 existing Alarm Evaluator, this new "Notification-driven Alarm Evaluator" is
764 aggregating and correlating different alarms which are then provided northbound
765 to the Consumer via the OpenStack "Alarm Notifier". The user/administrator can
766 register the alarm configuration via existing Ceilometer API [*]_. Thereby, he
767 can configure whether to set an alarm or not and where to send the alarms to.
769 **Implementation detail**
771 * The new "Notification-driven Alarm Evaluator" is part of Ceilometer.
772 * Most of the existing source code of the "Alarm Evaluator" can be re-used to
774 * No additional application logic is needed
775 * It will access the Ceilometer Databases just like the existing "Alarm
777 * Only the polling-based approach will be replaced by a listener for
778 "notifications" provided by the "Event Publisher for Alarm" on the Ceilometer
780 * No new interfaces have to be added to Ceilometer.
783 .. [*] https://etherpad.opnfv.org/p/doctor_bps
784 .. [*] https://wiki.openstack.org/wiki/Ceilometer/Alerting
786 Report host fault to update server state immediately (Nova) [*]_
787 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
789 **Problem statement:**
791 * Nova state change for failed or unreachable host is slow and does not reliably
792 state host is down or not. This might cause same server instance to run twice
793 if action taken to evacuate instance to another host.
794 * Nova state for server(s) on failed host will not change, but remains active
795 and running. This gives the user false information about server state.
796 * VIM northbound interface notification of host faults towards VNFM and NFVO
797 should be in line with OpenStack state. This fault notification is a Telco
798 requirement defined in ETSI and will be implemented by OPNFV Doctor project.
799 * Openstack user cannot make HA actions fast and reliably by trusting server
800 state and host state.
804 There needs to be a new API for Admin to state host is down. This API is used to
805 mark services running in host down to reflect the real situation.
807 Example on compute node is:
809 * When compute node is up and running:::
811 vm_state: activeand power_state: running
812 nova-compute state: up status: enabled
814 * When compute node goes down and new API is called to state host is down:::
816 vm_state: stopped power_state: shutdown
817 nova-compute state: down status: enabled
821 There is no attractive alternative to detect all different host faults than to
822 have an external tool to detect different host faults. For this kind of tool to
823 exist there needs to be new API in Nova to report fault. Currently there must be
824 some kind of workarounds implemented as cannot trust or get the states from
825 OpenStack fast enough.
827 .. [*] https://blueprints.launchpad.net/nova/+spec/update-server-state-immediately
832 This section lists some BPs related to Doctor, but proposed by drafters outside
835 pacemaker-servicegroup-driver [*]_
836 __________________________________
838 This BP will detect and report host down quite fast to OpenStack. This however
839 might not work properly for example when management network has some problem and
840 host reported faulty while VM still running there. This might lead to launching
841 same VM instance twice causing problems. Also NB IF message needs fault reason
842 and for that the source needs to be a tool that detects different kind of faults
843 as Doctor will be doing. Also this BP might need enhancement to change server
844 and service states correctly.
846 .. [*] https://blueprints.launchpad.net/nova/+spec/pacemaker-servicegroup-driver
849 vim: set tabstop=4 expandtab textwidth=80: