)]}'
{"specs/newton/amqp-dispatch-router.rst":[{"author":{"_account_id":2813,"name":"Mehdi Abaakouk","email":"sileht@sileht.net","username":"sileht"},"change_message_id":"1851de639779ab2409498272e95cbbd72ab4cb6b","unresolved":false,"context_lines":[{"line_number":452,"context_line":"Re-delivery will not be a goal of the Newton release.  The driver will"},{"line_number":453,"context_line":"simply treat the reception of a RELEASED or MODIFIED disposition as a"},{"line_number":454,"context_line":"message delivery failure.  This may be addressed in a different"},{"line_number":455,"context_line":"fashion in a future release."},{"line_number":456,"context_line":""},{"line_number":457,"context_line":""},{"line_number":458,"context_line":"Alternatives"}],"source_content_type":"text/x-rst","patch_set":1,"id":"dab17558_432406c4","line":455,"updated":"2016-05-10 15:30:46.000000000","message":"Perhaps we should just not implements the notification API for this drivers. Kafka and rabbitmq are more appropriate for notification API.\n\nNote for example, notification API means that Notification client can die, but we do not lost already produced messages. Some applications rely on this for monitoring/metering/billing/...\n\nWith the AMQP1.0 dispatch router system, we can\u0027t ensure that due to the nature of no\u0027broker approach.\n\nSo oslo.messaging have to store the messages somewhere itself to resend them later.\n\nZMQ have the same issue.","commit_id":"9bb4513f0b314f4bf1a67440d3c04d210dece4d8"},{"author":{"_account_id":8770,"name":"Ken Giusti","email":"kgiusti@gmail.com","username":"kgiusti"},"change_message_id":"4b1d9b5ccd8c05d7855a5fd89a28679c216a37c4","unresolved":false,"context_lines":[{"line_number":452,"context_line":"Re-delivery will not be a goal of the Newton release.  The driver will"},{"line_number":453,"context_line":"simply treat the reception of a RELEASED or MODIFIED disposition as a"},{"line_number":454,"context_line":"message delivery failure.  This may be addressed in a different"},{"line_number":455,"context_line":"fashion in a future release."},{"line_number":456,"context_line":""},{"line_number":457,"context_line":""},{"line_number":458,"context_line":"Alternatives"}],"source_content_type":"text/x-rst","patch_set":1,"id":"dab17558_adfd94b2","line":455,"in_reply_to":"dab17558_432406c4","updated":"2016-05-11 19:16:01.000000000","message":"Good point - for newton the notification api should raise a not implemented exception when called.  I\u0027ll update the spec to clarify that point.\n\nI would still like to plan for having the option of supporting Notifications by broker(s) that are integrated into the router network.  At least not design-out that capability out.  I suspect that there is value - in a distributed cloud - of having the ability to partition notification traffic across multiple brokers.\n\nThose brokers will of course support the same durability guarantees a directly-connected broker.  Notifications would never be stored by the driver as long as a broker is accessible.","commit_id":"9bb4513f0b314f4bf1a67440d3c04d210dece4d8"},{"author":{"_account_id":6159,"name":"Flavio Percoco Premoli","display_name":"flaper87","email":"flavio.percoco@flyrlabs.com","username":"flaper87"},"change_message_id":"04cf03f20fe1765d7304c3730967d4f94c77d0b5","unresolved":false,"context_lines":[{"line_number":5,"context_line":"https://blueprints.launchpad.net/oslo?searchtext\u003damqp-dispatch-router"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"This specification proposes changes to the existing AMQP 1.0 driver"},{"line_number":8,"context_line":"that will provide support for the Qpid Dispatch Router message bus as"},{"line_number":9,"context_line":"an RPC transport.  The Notification pattern will not be supported in"},{"line_number":10,"context_line":"the Newton timeframe and will be introduced in a later release."},{"line_number":11,"context_line":"The blueprint for the original driver can be found here [1]_ and the"}],"source_content_type":"text/x-rst","patch_set":2,"id":"dab17558_0751a1d8","line":8,"range":{"start_line":8,"start_character":34,"end_line":8,"end_character":54},"updated":"2016-05-13 23:04:56.000000000","message":"To be honest, I\u0027m a bit concerned about the mention of Qpid Dispatch Router here. The amqp1 driver was created as a protocol driver (technology agnostic) and this spec seems to be proposing a change of directions by adapting it to a specific technology.\n\nAm I reading this correctly? If this is not the case, I believe we should probably rephrase this paragraph to avoid misunderstandings.","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":8770,"name":"Ken Giusti","email":"kgiusti@gmail.com","username":"kgiusti"},"change_message_id":"a9208a148e48a77899633aa8a039cd73d67331f7","unresolved":false,"context_lines":[{"line_number":5,"context_line":"https://blueprints.launchpad.net/oslo?searchtext\u003damqp-dispatch-router"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"This specification proposes changes to the existing AMQP 1.0 driver"},{"line_number":8,"context_line":"that will provide support for the Qpid Dispatch Router message bus as"},{"line_number":9,"context_line":"an RPC transport.  The Notification pattern will not be supported in"},{"line_number":10,"context_line":"the Newton timeframe and will be introduced in a later release."},{"line_number":11,"context_line":"The blueprint for the original driver can be found here [1]_ and the"}],"source_content_type":"text/x-rst","patch_set":2,"id":"bab6814e_8e685abc","line":8,"range":{"start_line":8,"start_character":34,"end_line":8,"end_character":54},"in_reply_to":"dab17558_0751a1d8","updated":"2016-05-18 13:35:44.000000000","message":"I see how you\u0027d draw that conclusion.  But don\u0027t worry - these changes are in no way specific to the Qpid Dispatch Router. We\u0027re basically augmenting the driver in two areas:\n\n1) a credit management scheme that will work better for non-queueing messaging backends\n\n2) add more semantics to the addresses used by the driver\n\nNeither should cause lock-in to the Qpid Dispatch router.  And to be explicit these changes will also work with qpidd and should work with ActiveMQ, though I doubt I\u0027ll get ActiveMQ tested before newton.\n\nI will update the spec to clarify that.","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":6159,"name":"Flavio Percoco Premoli","display_name":"flaper87","email":"flavio.percoco@flyrlabs.com","username":"flaper87"},"change_message_id":"04cf03f20fe1765d7304c3730967d4f94c77d0b5","unresolved":false,"context_lines":[{"line_number":9,"context_line":"an RPC transport.  The Notification pattern will not be supported in"},{"line_number":10,"context_line":"the Newton timeframe and will be introduced in a later release."},{"line_number":11,"context_line":"The blueprint for the original driver can be found here [1]_ and the"},{"line_number":12,"context_line":"original architecture is described in [2]_.  The changes described by"},{"line_number":13,"context_line":"this specification are a refactoring of the original design.  Support"},{"line_number":14,"context_line":"for the Qpid C++ Router (qpidd) will be preserved."},{"line_number":15,"context_line":""},{"line_number":16,"context_line":"Problem description"},{"line_number":17,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":2,"id":"dab17558_07688186","line":14,"range":{"start_line":12,"start_character":45,"end_line":14,"end_character":50},"updated":"2016-05-13 23:04:56.000000000","message":"This seems to confirm my suspect. What changes in the architecture are expected?","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":8770,"name":"Ken Giusti","email":"kgiusti@gmail.com","username":"kgiusti"},"change_message_id":"a9208a148e48a77899633aa8a039cd73d67331f7","unresolved":false,"context_lines":[{"line_number":9,"context_line":"an RPC transport.  The Notification pattern will not be supported in"},{"line_number":10,"context_line":"the Newton timeframe and will be introduced in a later release."},{"line_number":11,"context_line":"The blueprint for the original driver can be found here [1]_ and the"},{"line_number":12,"context_line":"original architecture is described in [2]_.  The changes described by"},{"line_number":13,"context_line":"this specification are a refactoring of the original design.  Support"},{"line_number":14,"context_line":"for the Qpid C++ Router (qpidd) will be preserved."},{"line_number":15,"context_line":""},{"line_number":16,"context_line":"Problem description"},{"line_number":17,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":2,"id":"bab6814e_ee8c8637","line":14,"range":{"start_line":12,"start_character":45,"end_line":14,"end_character":50},"in_reply_to":"dab17558_07688186","updated":"2016-05-18 13:35:44.000000000","message":"As commented above, however the document goes into some detail of these changes in the \"Proposed Changes\" section.","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":6159,"name":"Flavio Percoco Premoli","display_name":"flaper87","email":"flavio.percoco@flyrlabs.com","username":"flaper87"},"change_message_id":"04cf03f20fe1765d7304c3730967d4f94c77d0b5","unresolved":false,"context_lines":[{"line_number":16,"context_line":"Problem description"},{"line_number":17,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":18,"context_line":""},{"line_number":19,"context_line":"The Qpid C++ Broker was the only messaging service that supported the"},{"line_number":20,"context_line":"AMQP 1.0 protocol when the original driver was developed.  As such,"},{"line_number":21,"context_line":"the original driver was designed and tested against a broker"},{"line_number":22,"context_line":"intermediary.  An alternative to the traditional broker-centric messaging architecture"},{"line_number":23,"context_line":"has become available since the driver was originally developed.  The Qpid Dispatch Router provides a"},{"line_number":24,"context_line":"decentralized, highly redundant, scalable messaging network that is"},{"line_number":25,"context_line":"quite different from the traditional centralized single-broker model."},{"line_number":26,"context_line":"This specification will describe the changes to the existing AMQP 1.0"},{"line_number":27,"context_line":"driver necessary for using Oslo Messaging services across a dispatch router"},{"line_number":28,"context_line":"network."},{"line_number":29,"context_line":""},{"line_number":30,"context_line":"The Qpid Dispatch Router is a stateless *message switch* that routes"},{"line_number":31,"context_line":"messages to their destination using a shortest path algorithm.  Think"}],"source_content_type":"text/x-rst","patch_set":2,"id":"dab17558_67d34d39","line":28,"range":{"start_line":19,"start_character":0,"end_line":28,"end_character":8},"updated":"2016-05-13 23:04:56.000000000","message":"Picking a technology that will be used for testing this driver is great but (sorry if I\u0027m just repeating myself over and over) this seems to suggest the driver will just work with Qpid Dispatch.","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":8770,"name":"Ken Giusti","email":"kgiusti@gmail.com","username":"kgiusti"},"change_message_id":"a9208a148e48a77899633aa8a039cd73d67331f7","unresolved":false,"context_lines":[{"line_number":16,"context_line":"Problem description"},{"line_number":17,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":18,"context_line":""},{"line_number":19,"context_line":"The Qpid C++ Broker was the only messaging service that supported the"},{"line_number":20,"context_line":"AMQP 1.0 protocol when the original driver was developed.  As such,"},{"line_number":21,"context_line":"the original driver was designed and tested against a broker"},{"line_number":22,"context_line":"intermediary.  An alternative to the traditional broker-centric messaging architecture"},{"line_number":23,"context_line":"has become available since the driver was originally developed.  The Qpid Dispatch Router provides a"},{"line_number":24,"context_line":"decentralized, highly redundant, scalable messaging network that is"},{"line_number":25,"context_line":"quite different from the traditional centralized single-broker model."},{"line_number":26,"context_line":"This specification will describe the changes to the existing AMQP 1.0"},{"line_number":27,"context_line":"driver necessary for using Oslo Messaging services across a dispatch router"},{"line_number":28,"context_line":"network."},{"line_number":29,"context_line":""},{"line_number":30,"context_line":"The Qpid Dispatch Router is a stateless *message switch* that routes"},{"line_number":31,"context_line":"messages to their destination using a shortest path algorithm.  Think"}],"source_content_type":"text/x-rst","patch_set":2,"id":"bab6814e_4e267269","line":28,"range":{"start_line":19,"start_character":0,"end_line":28,"end_character":8},"in_reply_to":"dab17558_67d34d39","updated":"2016-05-18 13:35:44.000000000","message":"Ok, I\u0027ll clarify that.","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":20523,"name":"Andy Smith","email":"ansmith@redhat.com","username":"ajssmith"},"change_message_id":"f156e82f2a9cc5370ce4f33f10f7db48ea005515","unresolved":false,"context_lines":[{"line_number":33,"context_line":"of IP address.  The router is **not a broker**.  It differs from broker"},{"line_number":34,"context_line":"in the following significant ways:"},{"line_number":35,"context_line":""},{"line_number":36,"context_line":"* it is deployed with other routers as a network"},{"line_number":37,"context_line":""},{"line_number":38,"context_line":"* the router is queue-less"},{"line_number":39,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"dab17558_cb180087","line":36,"range":{"start_line":36,"start_character":28,"end_line":36,"end_character":48},"updated":"2016-05-16 17:48:26.000000000","message":"The phrases \"routers as a network\" and \"routed network\" may be overloaded terms for certain stakeholders. Would \"messaging interconnect\" work as a general substitution?","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":8770,"name":"Ken Giusti","email":"kgiusti@gmail.com","username":"kgiusti"},"change_message_id":"a9208a148e48a77899633aa8a039cd73d67331f7","unresolved":false,"context_lines":[{"line_number":33,"context_line":"of IP address.  The router is **not a broker**.  It differs from broker"},{"line_number":34,"context_line":"in the following significant ways:"},{"line_number":35,"context_line":""},{"line_number":36,"context_line":"* it is deployed with other routers as a network"},{"line_number":37,"context_line":""},{"line_number":38,"context_line":"* the router is queue-less"},{"line_number":39,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"bab6814e_cebee2fe","line":36,"range":{"start_line":36,"start_character":28,"end_line":36,"end_character":48},"in_reply_to":"dab17558_cb180087","updated":"2016-05-18 13:35:44.000000000","message":"In this particular instance I\u0027d rather keep \u0027routers as a network\u0027, but I agreed \u0027messaging interconnect\u0027 is less vague that \u0027routed network\u0027 - which could be interpreted as IP routing.\n\nI will introduce that term and replace \u0027routed network\u0027 as appropriate.","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":20523,"name":"Andy Smith","email":"ansmith@redhat.com","username":"ajssmith"},"change_message_id":"f156e82f2a9cc5370ce4f33f10f7db48ea005515","unresolved":false,"context_lines":[{"line_number":61,"context_line":"subscriber which consumes the message."},{"line_number":62,"context_line":""},{"line_number":63,"context_line":"Note well that when a message transits a router on its way to its"},{"line_number":64,"context_line":"final destination no queueing is done.  To be specific, none of the"},{"line_number":65,"context_line":"routers along the path claim ownership of the message - i.e. the"},{"line_number":66,"context_line":"routers do not acknowledge the message.  Instead the ack from the"},{"line_number":67,"context_line":"consuming client is propagated back through the network to the origin"}],"source_content_type":"text/x-rst","patch_set":2,"id":"dab17558_ebace4a0","line":64,"updated":"2016-05-16 17:48:26.000000000","message":"It may be helpful to define queueing versus buffering where (1) queueing implies \"taking\" ownership and (2) buffering is what happens to a message on router transit.  This may help the \"Credit Management\" narrative below.","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":8770,"name":"Ken Giusti","email":"kgiusti@gmail.com","username":"kgiusti"},"change_message_id":"a9208a148e48a77899633aa8a039cd73d67331f7","unresolved":false,"context_lines":[{"line_number":61,"context_line":"subscriber which consumes the message."},{"line_number":62,"context_line":""},{"line_number":63,"context_line":"Note well that when a message transits a router on its way to its"},{"line_number":64,"context_line":"final destination no queueing is done.  To be specific, none of the"},{"line_number":65,"context_line":"routers along the path claim ownership of the message - i.e. the"},{"line_number":66,"context_line":"routers do not acknowledge the message.  Instead the ack from the"},{"line_number":67,"context_line":"consuming client is propagated back through the network to the origin"}],"source_content_type":"text/x-rst","patch_set":2,"id":"bab6814e_ae739e4c","line":64,"in_reply_to":"dab17558_ebace4a0","updated":"2016-05-18 13:35:44.000000000","message":"Will do.","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":20523,"name":"Andy Smith","email":"ansmith@redhat.com","username":"ajssmith"},"change_message_id":"f156e82f2a9cc5370ce4f33f10f7db48ea005515","unresolved":false,"context_lines":[{"line_number":151,"context_line":"Router D.  In this example notice that only those routers along the"},{"line_number":152,"context_line":"shortest path between two communicating endpoints are involved in the"},{"line_number":153,"context_line":"message transfer.  This provides for a degree of parallelism not"},{"line_number":154,"context_line":"possible when using a single broker backend."},{"line_number":155,"context_line":""},{"line_number":156,"context_line":"More information about the router can be found here [3]_."},{"line_number":157,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"dab17558_2b5f0be6","line":154,"updated":"2016-05-16 17:48:26.000000000","message":"Example would be enhanced with inclusion of link(AC) failure resulting in RPC Caller3 to Server1 message transit through Router B.","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":8770,"name":"Ken Giusti","email":"kgiusti@gmail.com","username":"kgiusti"},"change_message_id":"a9208a148e48a77899633aa8a039cd73d67331f7","unresolved":false,"context_lines":[{"line_number":151,"context_line":"Router D.  In this example notice that only those routers along the"},{"line_number":152,"context_line":"shortest path between two communicating endpoints are involved in the"},{"line_number":153,"context_line":"message transfer.  This provides for a degree of parallelism not"},{"line_number":154,"context_line":"possible when using a single broker backend."},{"line_number":155,"context_line":""},{"line_number":156,"context_line":"More information about the router can be found here [3]_."},{"line_number":157,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"bab6814e_4e8312cd","line":154,"in_reply_to":"dab17558_2b5f0be6","updated":"2016-05-18 13:35:44.000000000","message":"Agreed.","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":6159,"name":"Flavio Percoco Premoli","display_name":"flaper87","email":"flavio.percoco@flyrlabs.com","username":"flaper87"},"change_message_id":"04cf03f20fe1765d7304c3730967d4f94c77d0b5","unresolved":false,"context_lines":[{"line_number":159,"context_line":"Proposed change"},{"line_number":160,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":161,"context_line":""},{"line_number":162,"context_line":"Three issues need to be resolved in the driver in order to support RPC"},{"line_number":163,"context_line":"over a routed messaging network:"},{"line_number":164,"context_line":""},{"line_number":165,"context_line":"* optimal message addressing"},{"line_number":166,"context_line":"* credit management"}],"source_content_type":"text/x-rst","patch_set":2,"id":"dab17558_47e0c9cc","line":163,"range":{"start_line":162,"start_character":0,"end_line":163,"end_character":32},"updated":"2016-05-13 23:04:56.000000000","message":"Spec wise, this seems to be what we actually want to achieve. Move the driver out of the assumption there\u0027s a broker on the other side, assuming that assumption ever existed.\n\nThe above would bring the driver closer to what it was originally meant to do, right?\n\nAm I missing something here?","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":8770,"name":"Ken Giusti","email":"kgiusti@gmail.com","username":"kgiusti"},"change_message_id":"a9208a148e48a77899633aa8a039cd73d67331f7","unresolved":false,"context_lines":[{"line_number":159,"context_line":"Proposed change"},{"line_number":160,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":161,"context_line":""},{"line_number":162,"context_line":"Three issues need to be resolved in the driver in order to support RPC"},{"line_number":163,"context_line":"over a routed messaging network:"},{"line_number":164,"context_line":""},{"line_number":165,"context_line":"* optimal message addressing"},{"line_number":166,"context_line":"* credit management"}],"source_content_type":"text/x-rst","patch_set":2,"id":"bab6814e_ccdd4b72","line":163,"range":{"start_line":162,"start_character":0,"end_line":163,"end_character":32},"in_reply_to":"dab17558_47e0c9cc","updated":"2016-05-18 13:35:44.000000000","message":"It is exactly what we want to do - hence the \"Proposed change\" heading ;P\n\nThe reason the whole previous part of the spec went in to so much detail regarding the dispatch router is that \u0027messaging interconnect\u0027 is a new concept.  I feel the reader needs a very good understanding of how this backend works before they can judge the proposed changes in light of why the changes are actually needed.\n\nAdmittedly this spec is longer than usual and contains more detail about components out side of the driver implementation, but I think that detail is necessary in order to understand the architecture as a whole.","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":20523,"name":"Andy Smith","email":"ansmith@redhat.com","username":"ajssmith"},"change_message_id":"f156e82f2a9cc5370ce4f33f10f7db48ea005515","unresolved":false,"context_lines":[{"line_number":213,"context_line":"* fanout"},{"line_number":214,"context_line":"* direct reply"},{"line_number":215,"context_line":""},{"line_number":216,"context_line":"The following address syntax is proposed for the first three of the"},{"line_number":217,"context_line":"above patterns respectively::"},{"line_number":218,"context_line":""},{"line_number":219,"context_line":"  openstack.org/om/rpc/unicast/$EXCHANGE/$TOPIC/$SERVER"}],"source_content_type":"text/x-rst","patch_set":2,"id":"dab17558_eb7d046e","line":216,"updated":"2016-05-16 17:48:26.000000000","message":"Would building tags into a reverse-dns format to identify the type hierarchy provide an extensible pattern like the following:\n\norg.openstack.om.rpc.unicast/$EXCHANGE/$TOPIC/$SERVER\norg.openstack.om.rpc.anycast/$EXCHANGE/$TOPIC\norg.openstack.om.rpc.multicast/$EXCHANGE/$TOPIC","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":8770,"name":"Ken Giusti","email":"kgiusti@gmail.com","username":"kgiusti"},"change_message_id":"a9208a148e48a77899633aa8a039cd73d67331f7","unresolved":false,"context_lines":[{"line_number":213,"context_line":"* fanout"},{"line_number":214,"context_line":"* direct reply"},{"line_number":215,"context_line":""},{"line_number":216,"context_line":"The following address syntax is proposed for the first three of the"},{"line_number":217,"context_line":"above patterns respectively::"},{"line_number":218,"context_line":""},{"line_number":219,"context_line":"  openstack.org/om/rpc/unicast/$EXCHANGE/$TOPIC/$SERVER"}],"source_content_type":"text/x-rst","patch_set":2,"id":"bab6814e_8c6d0307","line":216,"in_reply_to":"dab17558_eb7d046e","updated":"2016-05-18 13:35:44.000000000","message":"These addresses are functionally equivalent to the proposed format.  However the AMQP 1.0 addressing draft spec uses \u0027/\u0027 as a component separator, not \u0027.\u0027.  \n\nI\u0027d like to stick with the \u0027/\u0027 and more partitioning in the address simply because I suspect a router could optimize the lookup tree by breaking it up into address components.\n\nI don\u0027t think that\u0027s the case for the dispatch router, but other implementations may be more pedantic with respect to the addressing spec.","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":6159,"name":"Flavio Percoco Premoli","display_name":"flaper87","email":"flavio.percoco@flyrlabs.com","username":"flaper87"},"change_message_id":"04cf03f20fe1765d7304c3730967d4f94c77d0b5","unresolved":false,"context_lines":[{"line_number":570,"context_line":"The default values for these options will be determined as the driver"},{"line_number":571,"context_line":"is tuned for performance.  These items include:"},{"line_number":572,"context_line":""},{"line_number":573,"context_line":"* Credit limit for Reply subscription"},{"line_number":574,"context_line":"* Credit limit for Server subscription"},{"line_number":575,"context_line":"* Credit limit for Topic subscription"},{"line_number":576,"context_line":"* Credit limit for Fanout subscription"},{"line_number":577,"context_line":"* default RPC call timeout  (TBD: may already exist?)"},{"line_number":578,"context_line":"* default cast timeout"},{"line_number":579,"context_line":"* default fanout timeout"}],"source_content_type":"text/x-rst","patch_set":2,"id":"dab17558_27e665b4","line":576,"range":{"start_line":573,"start_character":0,"end_line":576,"end_character":38},"updated":"2016-05-13 23:04:56.000000000","message":"Is this something that really needs to be configurable?\n\nCan we specify the default values here as well?","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":8770,"name":"Ken Giusti","email":"kgiusti@gmail.com","username":"kgiusti"},"change_message_id":"a9208a148e48a77899633aa8a039cd73d67331f7","unresolved":false,"context_lines":[{"line_number":570,"context_line":"The default values for these options will be determined as the driver"},{"line_number":571,"context_line":"is tuned for performance.  These items include:"},{"line_number":572,"context_line":""},{"line_number":573,"context_line":"* Credit limit for Reply subscription"},{"line_number":574,"context_line":"* Credit limit for Server subscription"},{"line_number":575,"context_line":"* Credit limit for Topic subscription"},{"line_number":576,"context_line":"* Credit limit for Fanout subscription"},{"line_number":577,"context_line":"* default RPC call timeout  (TBD: may already exist?)"},{"line_number":578,"context_line":"* default cast timeout"},{"line_number":579,"context_line":"* default fanout timeout"}],"source_content_type":"text/x-rst","patch_set":2,"id":"bab6814e_6ff891e6","line":576,"range":{"start_line":573,"start_character":0,"end_line":576,"end_character":38},"in_reply_to":"dab17558_27e665b4","updated":"2016-05-18 13:35:44.000000000","message":"It needs to be configurable because I doubt my default values will be optimal in every use case.  Recently the dispatch router made its credit limits configurable for this very reason - the default settings didn\u0027t perform well under higher loads.\n\nI will add the defaults.","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":6159,"name":"Flavio Percoco Premoli","display_name":"flaper87","email":"flavio.percoco@flyrlabs.com","username":"flaper87"},"change_message_id":"04cf03f20fe1765d7304c3730967d4f94c77d0b5","unresolved":false,"context_lines":[{"line_number":574,"context_line":"* Credit limit for Server subscription"},{"line_number":575,"context_line":"* Credit limit for Topic subscription"},{"line_number":576,"context_line":"* Credit limit for Fanout subscription"},{"line_number":577,"context_line":"* default RPC call timeout  (TBD: may already exist?)"},{"line_number":578,"context_line":"* default cast timeout"},{"line_number":579,"context_line":"* default fanout timeout"},{"line_number":580,"context_line":"* default reply timeout"}],"source_content_type":"text/x-rst","patch_set":2,"id":"dab17558_c772f9e2","line":577,"range":{"start_line":577,"start_character":28,"end_line":577,"end_character":52},"updated":"2016-05-13 23:04:56.000000000","message":"There\u0027s an `rpc_response_timeout` option, which determines how long a client should wait for a reply.","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"},{"author":{"_account_id":8770,"name":"Ken Giusti","email":"kgiusti@gmail.com","username":"kgiusti"},"change_message_id":"a9208a148e48a77899633aa8a039cd73d67331f7","unresolved":false,"context_lines":[{"line_number":574,"context_line":"* Credit limit for Server subscription"},{"line_number":575,"context_line":"* Credit limit for Topic subscription"},{"line_number":576,"context_line":"* Credit limit for Fanout subscription"},{"line_number":577,"context_line":"* default RPC call timeout  (TBD: may already exist?)"},{"line_number":578,"context_line":"* default cast timeout"},{"line_number":579,"context_line":"* default fanout timeout"},{"line_number":580,"context_line":"* default reply timeout"}],"source_content_type":"text/x-rst","patch_set":2,"id":"bab6814e_0f0465be","line":577,"range":{"start_line":577,"start_character":28,"end_line":577,"end_character":52},"in_reply_to":"dab17558_c772f9e2","updated":"2016-05-18 13:35:44.000000000","message":"Great - the driver will use that for the RPC round trip instead of defining a driver-specific one.","commit_id":"22310d42004d230e0c27724ba060c9f72fff9cbc"}]}
