)]}'
{"specs/mitaka/approved/disk-ram-monitor-plugin.rst":[{"author":{"_account_id":1297,"name":"Joshua Harlow","email":"jxharlow@godaddy.com","username":"harlowja"},"change_message_id":"90762eb866b6886708795f4b7651952701e7ebea","unresolved":false,"context_lines":[{"line_number":14,"context_line":"in available. There are various occasion when the admin would like to know"},{"line_number":15,"context_line":"the actual utilization of disk and memory on the compute. With this feature"},{"line_number":16,"context_line":"we would like to introduce two new monitor plugins one for disk and other"},{"line_number":17,"context_line":"for memory to report the actual usage statistics. These statistics will be"},{"line_number":18,"context_line":"dumped into nove.compute_nodes tables as part of the metrics column and can"},{"line_number":19,"context_line":"be fetched from the API. Also the scheduler changes to understand these"},{"line_number":20,"context_line":"metrics will be included as part of this feature."}],"source_content_type":"text/x-rst","patch_set":1,"id":"7a740942_9d93bd5a","line":17,"updated":"2015-12-09 08:11:36.000000000","message":"Hmmm, interesting, does ceilometer/telementry projec already gather this data? Maybe useful to see what these projects already gather, and uses there information in the scheduler instead... Avoids creating new \u0027monitor\u0027 plugins and could be a nice integration point, probably something worth investigating imho....\n\nhttps://wiki.openstack.org/wiki/Telemetry","commit_id":"4905c78d6830fa38fd9fcd9b48ca5e4f958f078b"},{"author":{"_account_id":10263,"name":"Vilobh Meshram","email":"vilobhmm@yahoo-inc.com","username":"vilobhmm"},"change_message_id":"120d79672d77a7ec864d694d9ad790c6f17a56ca","unresolved":false,"context_lines":[{"line_number":14,"context_line":"in available. There are various occasion when the admin would like to know"},{"line_number":15,"context_line":"the actual utilization of disk and memory on the compute. With this feature"},{"line_number":16,"context_line":"we would like to introduce two new monitor plugins one for disk and other"},{"line_number":17,"context_line":"for memory to report the actual usage statistics. These statistics will be"},{"line_number":18,"context_line":"dumped into nove.compute_nodes tables as part of the metrics column and can"},{"line_number":19,"context_line":"be fetched from the API. Also the scheduler changes to understand these"},{"line_number":20,"context_line":"metrics will be included as part of this feature."}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a710552_c22ab02c","line":17,"in_reply_to":"7a740942_53002f02","updated":"2015-12-14 23:38:47.000000000","message":"@Josh : I agree telemetry is also collecting utilization data but IMHO not every private cloud deployments will be running ceilometer or telemetry. So adding something locally for nova\n\nhttps://github.com/openstack/nova/tree/master/nova/compute/monitors/\n\nshould help. Thoughts ?\n\n+1 Joe","commit_id":"4905c78d6830fa38fd9fcd9b48ca5e4f958f078b"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"c4af4f149b5ac2b87821c43e9795ccc165ce85d0","unresolved":false,"context_lines":[{"line_number":14,"context_line":"in available. There are various occasion when the admin would like to know"},{"line_number":15,"context_line":"the actual utilization of disk and memory on the compute. With this feature"},{"line_number":16,"context_line":"we would like to introduce two new monitor plugins one for disk and other"},{"line_number":17,"context_line":"for memory to report the actual usage statistics. These statistics will be"},{"line_number":18,"context_line":"dumped into nove.compute_nodes tables as part of the metrics column and can"},{"line_number":19,"context_line":"be fetched from the API. Also the scheduler changes to understand these"},{"line_number":20,"context_line":"metrics will be included as part of this feature."}],"source_content_type":"text/x-rst","patch_set":1,"id":"7a740942_53002f02","line":17,"in_reply_to":"7a740942_9d93bd5a","updated":"2015-12-12 07:55:25.000000000","message":"Even if telemetry gathers, I don\u0027t think it\u0027s a reason that nova metrics couldn\u0027t as well since they\u0027re sort of different.  Nova metrics is very \u0027light weight\u0027 and intended for \u0027fast\u0027 access (e.g., for scheduling decisions) where querying telemetry data isn\u0027t really feasible/appropriate for \u0027real time\u0027 types of decisions.\n\nPerfect example is host CPU utilization - nova metrics has that, and so does telemetry (IIRC).","commit_id":"4905c78d6830fa38fd9fcd9b48ca5e4f958f078b"},{"author":{"_account_id":1297,"name":"Joshua Harlow","email":"jxharlow@godaddy.com","username":"harlowja"},"change_message_id":"90762eb866b6886708795f4b7651952701e7ebea","unresolved":false,"context_lines":[{"line_number":44,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":45,"context_line":"Proposed change is to write new monitor plugin for disk and memory to fetch"},{"line_number":46,"context_line":"the actual usage data from the compute node. The collected data will be"},{"line_number":47,"context_line":"collected by the periodic run of Resource Tracker and later dumped into"},{"line_number":48,"context_line":"nova.compute_nodes table metrics column. The actual usage information"},{"line_number":49,"context_line":"can be a very useful information at the nova API layer."},{"line_number":50,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"7a740942_5da795a6","line":47,"updated":"2015-12-09 08:11:36.000000000","message":"Sounds like ceilometer pollsters a little...\n\nhttp://docs.openstack.org/developer/ceilometer/architecture.html#gathering-the-data","commit_id":"4905c78d6830fa38fd9fcd9b48ca5e4f958f078b"},{"author":{"_account_id":10263,"name":"Vilobh Meshram","email":"vilobhmm@yahoo-inc.com","username":"vilobhmm"},"change_message_id":"120d79672d77a7ec864d694d9ad790c6f17a56ca","unresolved":false,"context_lines":[{"line_number":44,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":45,"context_line":"Proposed change is to write new monitor plugin for disk and memory to fetch"},{"line_number":46,"context_line":"the actual usage data from the compute node. The collected data will be"},{"line_number":47,"context_line":"collected by the periodic run of Resource Tracker and later dumped into"},{"line_number":48,"context_line":"nova.compute_nodes table metrics column. The actual usage information"},{"line_number":49,"context_line":"can be a very useful information at the nova API layer."},{"line_number":50,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a710552_c2a3904e","line":47,"in_reply_to":"7a740942_4e172445","updated":"2015-12-14 23:38:47.000000000","message":"The proposal here is to add new monitor for disk and memory as done for cpu https://github.com/openstack/nova/tree/master/nova/compute/monitors/cpu to collect the data for disk/memory usage and then dump it into nova db to be used by nova scheduler for making decisions based on utilization.","commit_id":"4905c78d6830fa38fd9fcd9b48ca5e4f958f078b"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"c4af4f149b5ac2b87821c43e9795ccc165ce85d0","unresolved":false,"context_lines":[{"line_number":44,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":45,"context_line":"Proposed change is to write new monitor plugin for disk and memory to fetch"},{"line_number":46,"context_line":"the actual usage data from the compute node. The collected data will be"},{"line_number":47,"context_line":"collected by the periodic run of Resource Tracker and later dumped into"},{"line_number":48,"context_line":"nova.compute_nodes table metrics column. The actual usage information"},{"line_number":49,"context_line":"can be a very useful information at the nova API layer."},{"line_number":50,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"7a740942_4e172445","line":47,"in_reply_to":"7a740942_5da795a6","updated":"2015-12-12 07:55:25.000000000","message":"Similar to above, I think this proposal is fine since it\u0027s just adding a new monitor to collect the new data metrics using the \"already existing\" metric framework.","commit_id":"4905c78d6830fa38fd9fcd9b48ca5e4f958f078b"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"c4af4f149b5ac2b87821c43e9795ccc165ce85d0","unresolved":false,"context_lines":[{"line_number":110,"context_line":"Work Items"},{"line_number":111,"context_line":"----------"},{"line_number":112,"context_line":""},{"line_number":113,"context_line":"1. Introduce Disk Monitor plugin."},{"line_number":114,"context_line":"2. Introduce Memory Monitor plugin."},{"line_number":115,"context_line":"3. Unit tests for Disk Monitor plugin."},{"line_number":116,"context_line":"4. Unit tests for Memory Monitor plugin."}],"source_content_type":"text/x-rst","patch_set":1,"id":"7a740942_ae1c8866","line":113,"updated":"2015-12-12 07:55:25.000000000","message":"Can we clarify some more details of the disk monitor plugin?  For example, does this represent the root disk only, all disks, etc.?  I assume each metric (if more than one) would represent an individual disk\u0027s utilization %?","commit_id":"4905c78d6830fa38fd9fcd9b48ca5e4f958f078b"},{"author":{"_account_id":10263,"name":"Vilobh Meshram","email":"vilobhmm@yahoo-inc.com","username":"vilobhmm"},"change_message_id":"b9431f97f3ceb37d725c46f01c44b44cb120a8af","unresolved":false,"context_lines":[{"line_number":110,"context_line":"Work Items"},{"line_number":111,"context_line":"----------"},{"line_number":112,"context_line":""},{"line_number":113,"context_line":"1. Introduce Disk Monitor plugin."},{"line_number":114,"context_line":"2. Introduce Memory Monitor plugin."},{"line_number":115,"context_line":"3. Unit tests for Disk Monitor plugin."},{"line_number":116,"context_line":"4. Unit tests for Memory Monitor plugin."}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a710552_e111e05e","line":113,"in_reply_to":"5a710552_81e25c9e","updated":"2015-12-14 23:55:47.000000000","message":"Agree with the mount point argument that its the mount point that gets written eventually. So do you plan to say it makes more sense to say it out that DiskMonitor plugin will dump the stats for the disk utilization for a particular mount point than for each individual disk. \n\nI would prefer going with #2 as it adds more value.\n\nAll mount points does not make sense IMHO. I think we should look for mount points with \n\n$instance_path\n$snapshot_path \n\nThoughts ?","commit_id":"4905c78d6830fa38fd9fcd9b48ca5e4f958f078b"},{"author":{"_account_id":10263,"name":"Vilobh Meshram","email":"vilobhmm@yahoo-inc.com","username":"vilobhmm"},"change_message_id":"beb9bf20626ac6ccadc559e4077f582a4c5b3123","unresolved":false,"context_lines":[{"line_number":110,"context_line":"Work Items"},{"line_number":111,"context_line":"----------"},{"line_number":112,"context_line":""},{"line_number":113,"context_line":"1. Introduce Disk Monitor plugin."},{"line_number":114,"context_line":"2. Introduce Memory Monitor plugin."},{"line_number":115,"context_line":"3. Unit tests for Disk Monitor plugin."},{"line_number":116,"context_line":"4. Unit tests for Memory Monitor plugin."}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a710552_ef2467b0","line":113,"in_reply_to":"5a710552_a46d2634","updated":"2015-12-15 00:29:53.000000000","message":"Just to be clear\n\n/home     /home/nova/instances (Mount Point)\n\n/var         /var/nova/snapshots (Mount Point)\n\nso we need to get utilization data for both mount points [/home/nova/instances and  /var/nova/snapshots]\n\nThis will be the main role of the Disk Monitor plugin.\n\nDoes this make sense ? Are we on same page here ?\n\nIf this is clear I will proceed ahead with  modifying the spec accordingly.","commit_id":"4905c78d6830fa38fd9fcd9b48ca5e4f958f078b"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"8c2eb0fbc52c414df24103cc4da4dcd9bf4814b3","unresolved":false,"context_lines":[{"line_number":110,"context_line":"Work Items"},{"line_number":111,"context_line":"----------"},{"line_number":112,"context_line":""},{"line_number":113,"context_line":"1. Introduce Disk Monitor plugin."},{"line_number":114,"context_line":"2. Introduce Memory Monitor plugin."},{"line_number":115,"context_line":"3. Unit tests for Disk Monitor plugin."},{"line_number":116,"context_line":"4. Unit tests for Memory Monitor plugin."}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a710552_a46d2634","line":113,"in_reply_to":"5a710552_e111e05e","updated":"2015-12-15 00:18:04.000000000","message":"When I said \"all mount points\", I probably wasn\u0027t clear.  I meant we\u0027d provide utilization for each *individual* mount point... not an aggregated view.","commit_id":"4905c78d6830fa38fd9fcd9b48ca5e4f958f078b"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"79f2c3244a8c9343e091b8c69df0fb5d2cfb7204","unresolved":false,"context_lines":[{"line_number":110,"context_line":"Work Items"},{"line_number":111,"context_line":"----------"},{"line_number":112,"context_line":""},{"line_number":113,"context_line":"1. Introduce Disk Monitor plugin."},{"line_number":114,"context_line":"2. Introduce Memory Monitor plugin."},{"line_number":115,"context_line":"3. Unit tests for Disk Monitor plugin."},{"line_number":116,"context_line":"4. Unit tests for Memory Monitor plugin."}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a710552_113f90b5","line":113,"in_reply_to":"5a710552_ef2467b0","updated":"2015-12-15 07:12:09.000000000","message":"Yes, I think that makes sense.  Appreciate it!","commit_id":"4905c78d6830fa38fd9fcd9b48ca5e4f958f078b"},{"author":{"_account_id":7664,"name":"Joe Cropper","email":"jwcroppe@us.ibm.com","username":"jwcroppe"},"change_message_id":"dd7012ba7da1772326bf55861366a2ed0a7ac8d6","unresolved":false,"context_lines":[{"line_number":110,"context_line":"Work Items"},{"line_number":111,"context_line":"----------"},{"line_number":112,"context_line":""},{"line_number":113,"context_line":"1. Introduce Disk Monitor plugin."},{"line_number":114,"context_line":"2. Introduce Memory Monitor plugin."},{"line_number":115,"context_line":"3. Unit tests for Disk Monitor plugin."},{"line_number":116,"context_line":"4. Unit tests for Memory Monitor plugin."}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a710552_81e25c9e","line":113,"in_reply_to":"5a710552_f62531ae","updated":"2015-12-14 23:48:34.000000000","message":"IMO, I think we\u0027d have utilization % for mount points... and somehow the metric name would indicate the mount point?\n\nFor example, from a scheduling perspective, someone would want to know which mount point to look at and its utilization across hosts since ultimately it\u0027s a mount point that gets \"written to\" ... e.g., that\u0027s why [df -h] shows you the mount points\u0027 utilization percentages.\n\nI think the spec needs to clarify that and also which mount point(s) it supports. I think it\u0027d either be (to start with) (1) root only; or (2) all mount points on the compute host.","commit_id":"4905c78d6830fa38fd9fcd9b48ca5e4f958f078b"},{"author":{"_account_id":10263,"name":"Vilobh Meshram","email":"vilobhmm@yahoo-inc.com","username":"vilobhmm"},"change_message_id":"120d79672d77a7ec864d694d9ad790c6f17a56ca","unresolved":false,"context_lines":[{"line_number":110,"context_line":"Work Items"},{"line_number":111,"context_line":"----------"},{"line_number":112,"context_line":""},{"line_number":113,"context_line":"1. Introduce Disk Monitor plugin."},{"line_number":114,"context_line":"2. Introduce Memory Monitor plugin."},{"line_number":115,"context_line":"3. Unit tests for Disk Monitor plugin."},{"line_number":116,"context_line":"4. Unit tests for Memory Monitor plugin."}],"source_content_type":"text/x-rst","patch_set":1,"id":"5a710552_f62531ae","line":113,"in_reply_to":"7a740942_ae1c8866","updated":"2015-12-14 23:38:47.000000000","message":"Yes ideally it would represent an individual disk\u0027s utilization. Gathering individual disk utilization vs gathering utilization for a mount point (which can be a logical gathering of a share (grouping of disk)) which ones makes more sense needs to be decided.\n\nIMHO getting an individual disk\u0027s utilization makes more sense.","commit_id":"4905c78d6830fa38fd9fcd9b48ca5e4f958f078b"}]}