)]}'
{"specs/juno/central-agent-partitioning.rst":[{"author":{"_account_id":4491,"name":"Lianhao Lu","email":"llh_misc@outlook.com","username":"lianhao-lu"},"change_message_id":"88bc8a9cb37c13c400ec55d75a2a49f596353c78","unresolved":false,"context_lines":[{"line_number":132,"context_line":"    partition_coordinator.start()"},{"line_number":133,"context_line":""},{"line_number":134,"context_line":"    every polling_interval:"},{"line_number":135,"context_line":"        all_resources \u003d discover_resources()"},{"line_number":136,"context_line":"        my_resource \u003d partition_coordinator.get_my_subset(all_resources)"},{"line_number":137,"context_line":"        for r in my_resources:"},{"line_number":138,"context_line":"            poll(r)"}],"source_content_type":"text/x-rst","patch_set":1,"id":"baada198_3bd9d128","line":135,"updated":"2014-08-07 06:04:15.000000000","message":"Does this means that all the central agents are required to use the same pipeline configuration? at least the same pipeline source configuration?\n\nCurrently the discovery mechanism is per pipeline source, for e.g. if we have central agent A with pipeline source Sa and discoverer Da, and another central agent B with different source Sb and discoverer Db, what will happen grouping/coordinator in this case?","commit_id":"c2be32ab84ba35526314090d72fbfa624940de3f"},{"author":{"_account_id":2284,"name":"Eoghan Glynn","email":"eglynn@redhat.com","username":"eglynn"},"change_message_id":"495528f76daac7da2d0137149c70149aa873b68a","unresolved":false,"context_lines":[{"line_number":132,"context_line":"    partition_coordinator.start()"},{"line_number":133,"context_line":""},{"line_number":134,"context_line":"    every polling_interval:"},{"line_number":135,"context_line":"        all_resources \u003d discover_resources()"},{"line_number":136,"context_line":"        my_resource \u003d partition_coordinator.get_my_subset(all_resources)"},{"line_number":137,"context_line":"        for r in my_resources:"},{"line_number":138,"context_line":"            poll(r)"}],"source_content_type":"text/x-rst","patch_set":1,"id":"baada198_af6cb808","line":135,"in_reply_to":"baada198_3bd9d128","updated":"2014-08-07 10:37:16.000000000","message":"As I understand it, consistent pipeline configurations would be required for the scaled-out central agents *if* we require that the resources polled by the central agent are each metered at a consistent cadence over their individual lifecycle.\n\nIn a sense, it\u0027s somewhat like the situation we currently have with the compute agent. It\u0027s possible, but inadvisable, to have different intervals configured for say the cpu_source in the pipeline.yaml files on two compute nodes. However if an instance is then migrated from hostA to hostB, suddenly the cadence of cpu_util metering for that instance is changed.\n\nAdmittedly instance migration is much less frequent than a resource hand-over between scaled-out central agents, but the principle is much the same I think.","commit_id":"c2be32ab84ba35526314090d72fbfa624940de3f"},{"author":{"_account_id":4491,"name":"Lianhao Lu","email":"llh_misc@outlook.com","username":"lianhao-lu"},"change_message_id":"da14e8ec18de2a857015b90749acb0bbd81039f3","unresolved":false,"context_lines":[{"line_number":132,"context_line":"    partition_coordinator.start()"},{"line_number":133,"context_line":""},{"line_number":134,"context_line":"    every polling_interval:"},{"line_number":135,"context_line":"        all_resources \u003d discover_resources()"},{"line_number":136,"context_line":"        my_resource \u003d partition_coordinator.get_my_subset(all_resources)"},{"line_number":137,"context_line":"        for r in my_resources:"},{"line_number":138,"context_line":"            poll(r)"}],"source_content_type":"text/x-rst","patch_set":1,"id":"baada198_f02be479","line":135,"in_reply_to":"baada198_af6cb808","updated":"2014-08-07 15:14:26.000000000","message":"I think some HP guys have mentioned that their usage model like different central agents polling on different things. I think it\u0027d be better that this bp wouldn\u0027t block their existing usage model.","commit_id":"c2be32ab84ba35526314090d72fbfa624940de3f"}]}