)]}'
{"specs/ocata/approved/prune-scheduler-cache-refresh.rst":[{"author":{"_account_id":13692,"name":"Roman Dobosz","email":"gryf73@gmail.com","username":"gryf"},"change_message_id":"37a5580ad41ab7aca241c6eae6c4aa0cc2c7c898","unresolved":false,"context_lines":[{"line_number":94,"context_line":"concurrent cache-refresh requests to one actual operation. The same result will"},{"line_number":95,"context_line":"be returned to all the waiting requests. It works in the following scenario:"},{"line_number":96,"context_line":""},{"line_number":97,"context_line":"  * When the host manager is querying host states for request A, and another"},{"line_number":98,"context_line":"    request B comes, the host manager won’t launch a second cache-refresh;"},{"line_number":99,"context_line":"    Instead, it simply reuses the first one and returns the same result to both"},{"line_number":100,"context_line":"    A and B."},{"line_number":101,"context_line":""},{"line_number":102,"context_line":"In this way, we can reduce the expensive cache-refresh queries to minimum while"},{"line_number":103,"context_line":"keeping scheduler host states fresh. And the cache-refresh can become much"}],"source_content_type":"text/x-rst","patch_set":1,"id":"bacf61ea_07c49e3d","line":100,"range":{"start_line":97,"start_character":0,"end_line":100,"end_character":12},"updated":"2016-08-01 17:30:25.000000000","message":"Does it mean, that cache lasts till no requests are made during query of the host state? In other words: the proposed change is to eliminate overlapping queries?","commit_id":"bf4970f24c5ee84802805b0f8ca15975df47e88a"},{"author":{"_account_id":15648,"name":"Yingxin","email":"yingxin.cheng@intel.com","username":"Yingxin"},"change_message_id":"cfc64f8d7202524b5c51c8ee8109daecf556ddcd","unresolved":false,"context_lines":[{"line_number":94,"context_line":"concurrent cache-refresh requests to one actual operation. The same result will"},{"line_number":95,"context_line":"be returned to all the waiting requests. It works in the following scenario:"},{"line_number":96,"context_line":""},{"line_number":97,"context_line":"  * When the host manager is querying host states for request A, and another"},{"line_number":98,"context_line":"    request B comes, the host manager won’t launch a second cache-refresh;"},{"line_number":99,"context_line":"    Instead, it simply reuses the first one and returns the same result to both"},{"line_number":100,"context_line":"    A and B."},{"line_number":101,"context_line":""},{"line_number":102,"context_line":"In this way, we can reduce the expensive cache-refresh queries to minimum while"},{"line_number":103,"context_line":"keeping scheduler host states fresh. And the cache-refresh can become much"}],"source_content_type":"text/x-rst","patch_set":1,"id":"bacf61ea_c32ee9a3","line":100,"range":{"start_line":97,"start_character":0,"end_line":100,"end_character":12},"in_reply_to":"bacf61ea_07c49e3d","updated":"2016-08-02 01:26:11.000000000","message":"Yes, you\u0027re correct.","commit_id":"bf4970f24c5ee84802805b0f8ca15975df47e88a"},{"author":{"_account_id":15648,"name":"Yingxin","email":"yingxin.cheng@intel.com","username":"Yingxin"},"change_message_id":"8535650099775793f8ae384d32fc80c6fe1e4d82","unresolved":false,"context_lines":[{"line_number":94,"context_line":"concurrent cache-refresh requests to one actual operation. The same result will"},{"line_number":95,"context_line":"be returned to all the waiting requests. It works in the following scenario:"},{"line_number":96,"context_line":""},{"line_number":97,"context_line":"  * When the host manager is querying host states for request A, and another"},{"line_number":98,"context_line":"    request B comes, the host manager won’t launch a second cache-refresh;"},{"line_number":99,"context_line":"    Instead, it simply reuses the first one and returns the same result to both"},{"line_number":100,"context_line":"    A and B."},{"line_number":101,"context_line":""},{"line_number":102,"context_line":"In this way, we can reduce the expensive cache-refresh queries to minimum while"},{"line_number":103,"context_line":"keeping scheduler host states fresh. And the cache-refresh can become much"}],"source_content_type":"text/x-rst","patch_set":1,"id":"bacf61ea_5e49f18d","line":100,"range":{"start_line":97,"start_character":0,"end_line":100,"end_character":12},"in_reply_to":"bacf61ea_38dba07d","updated":"2016-08-03 01:50:31.000000000","message":"The scheduler cache is stored in each scheduler instance independently. This spec won\u0027t change that fact to keep simple. \n\nAlthough it causes race conditions of decisions from multiple schedulers, but it is resolved by resource tracker claims and retries.","commit_id":"bf4970f24c5ee84802805b0f8ca15975df47e88a"},{"author":{"_account_id":13692,"name":"Roman Dobosz","email":"gryf73@gmail.com","username":"gryf"},"change_message_id":"bd94a23ae4ab226005ffdad091167245693de22d","unresolved":false,"context_lines":[{"line_number":94,"context_line":"concurrent cache-refresh requests to one actual operation. The same result will"},{"line_number":95,"context_line":"be returned to all the waiting requests. It works in the following scenario:"},{"line_number":96,"context_line":""},{"line_number":97,"context_line":"  * When the host manager is querying host states for request A, and another"},{"line_number":98,"context_line":"    request B comes, the host manager won’t launch a second cache-refresh;"},{"line_number":99,"context_line":"    Instead, it simply reuses the first one and returns the same result to both"},{"line_number":100,"context_line":"    A and B."},{"line_number":101,"context_line":""},{"line_number":102,"context_line":"In this way, we can reduce the expensive cache-refresh queries to minimum while"},{"line_number":103,"context_line":"keeping scheduler host states fresh. And the cache-refresh can become much"}],"source_content_type":"text/x-rst","patch_set":1,"id":"bacf61ea_b56c8337","line":100,"range":{"start_line":97,"start_character":0,"end_line":100,"end_character":12},"in_reply_to":"bacf61ea_5e49f18d","updated":"2016-08-03 07:03:44.000000000","message":"Ok, thanks, that makes sense.","commit_id":"bf4970f24c5ee84802805b0f8ca15975df47e88a"},{"author":{"_account_id":13692,"name":"Roman Dobosz","email":"gryf73@gmail.com","username":"gryf"},"change_message_id":"7ba38e7ec5c3e65f5327ae5fa3860b5fc12d4120","unresolved":false,"context_lines":[{"line_number":94,"context_line":"concurrent cache-refresh requests to one actual operation. The same result will"},{"line_number":95,"context_line":"be returned to all the waiting requests. It works in the following scenario:"},{"line_number":96,"context_line":""},{"line_number":97,"context_line":"  * When the host manager is querying host states for request A, and another"},{"line_number":98,"context_line":"    request B comes, the host manager won’t launch a second cache-refresh;"},{"line_number":99,"context_line":"    Instead, it simply reuses the first one and returns the same result to both"},{"line_number":100,"context_line":"    A and B."},{"line_number":101,"context_line":""},{"line_number":102,"context_line":"In this way, we can reduce the expensive cache-refresh queries to minimum while"},{"line_number":103,"context_line":"keeping scheduler host states fresh. And the cache-refresh can become much"}],"source_content_type":"text/x-rst","patch_set":1,"id":"bacf61ea_38dba07d","line":100,"range":{"start_line":97,"start_character":0,"end_line":100,"end_character":12},"in_reply_to":"bacf61ea_c32ee9a3","updated":"2016-08-02 10:58:46.000000000","message":"Ok, thanks. Another question would be - how it will behave in case of multiple schedulers? I assume, that there could be moment in time, that two different scheduler instances got request. Now, we would like to refresh hosts state and cache it. Where would be such cache stored? Or is this done for each scheduler instance independently?","commit_id":"bf4970f24c5ee84802805b0f8ca15975df47e88a"}]}