)]}'
{"keystone/cmd/doctor/caching.py":[{"author":{"_account_id":5046,"name":"Lance Bragstad","email":"lbragstad@redhat.com","username":"ldbragst"},"change_message_id":"5a6c5d44db6c6b37b8ec8f205cbedfb1e8644952","unresolved":false,"context_lines":[{"line_number":35,"context_line":"    return CONF.cache.enabled and CONF.cache.backend \u003d\u003d \u0027dogpile.cache.null\u0027"},{"line_number":36,"context_line":""},{"line_number":37,"context_line":""},{"line_number":38,"context_line":"def symptom_caching_debug_log_enabled():"},{"line_number":39,"context_line":"    \"\"\"Enabling caching log is not completely configured."},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"    Although debug level is enabled for Keystone, it\u0027s needed to enabled debug"}],"source_content_type":"text/x-python","patch_set":2,"id":"ff570b3c_6a9b6772","line":38,"updated":"2020-05-29 17:25:09.000000000","message":"I\u0027m thinking about the usability of this:\n\n- user is suspicious of a caching issue and wants to make sure keystone is using the cache\n- user runs doctor and notices they need to update the config\n- user updates the config\n- user re-runs doctor to make sure debug caching is configured properly\n- user bounces keystone and inspects logs\n\nIs that correct?\n\nAnother thing we could do pull the memcached string from keystone\u0027s CONF object and attempt to connect to it using dogpile directly. If the connection is successful, we don\u0027t emit the symptom. If it isn\u0027t successful, we tell the user that cache configuration is incorrect and they should go investigate. That would result in a flow like the following:\n\n- user is suspicious of a caching issue\n- user runs doctor and sees that keystone can\u0027t connect to memcached","commit_id":"772b807b22ff5c57cc05cc803a8096552f75bd89"},{"author":{"_account_id":8866,"name":"Raildo Mascena de Sousa Filho","email":"rmascena@redhat.com","username":"raildo"},"change_message_id":"d7098e264304ca001328476bbd4c4f785b059b05","unresolved":false,"context_lines":[{"line_number":35,"context_line":"    return CONF.cache.enabled and CONF.cache.backend \u003d\u003d \u0027dogpile.cache.null\u0027"},{"line_number":36,"context_line":""},{"line_number":37,"context_line":""},{"line_number":38,"context_line":"def symptom_caching_debug_log_enabled():"},{"line_number":39,"context_line":"    \"\"\"Enabling caching log is not completely configured."},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"    Although debug level is enabled for Keystone, it\u0027s needed to enabled debug"}],"source_content_type":"text/x-python","patch_set":2,"id":"ff570b3c_ea0ad776","line":38,"in_reply_to":"ff570b3c_6a9b6772","updated":"2020-05-29 17:34:27.000000000","message":"I\u0027m going for the first option that you described, trying to provide some more generic solution that can be used for other oslo.cache backends besides memcached.\n\nI\u0027m ok on making this more specific for memcached as you described, or maybe creating another symptom for memcached and keep this one here, thoughts?","commit_id":"772b807b22ff5c57cc05cc803a8096552f75bd89"},{"author":{"_account_id":5046,"name":"Lance Bragstad","email":"lbragstad@redhat.com","username":"ldbragst"},"change_message_id":"45415e6df0d89f2674f1aab24284166c3fa76b39","unresolved":false,"context_lines":[{"line_number":35,"context_line":"    return CONF.cache.enabled and CONF.cache.backend \u003d\u003d \u0027dogpile.cache.null\u0027"},{"line_number":36,"context_line":""},{"line_number":37,"context_line":""},{"line_number":38,"context_line":"def symptom_caching_debug_log_enabled():"},{"line_number":39,"context_line":"    \"\"\"Enabling caching log is not completely configured."},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"    Although debug level is enabled for Keystone, it\u0027s needed to enabled debug"}],"source_content_type":"text/x-python","patch_set":2,"id":"bf51134e_92ae1796","line":38,"in_reply_to":"ff570b3c_ea0ad776","updated":"2020-06-18 15:16:09.000000000","message":"My only hesitation about including assertions about debug logging is that the primary usecase for `keystone-manage doctor` is to tune a production system.\n\nI\u0027m not sure we\u0027d want to recommend debug logging in production since dogpile and keystone will be extremely noisy and possibly cause issues with disk space.","commit_id":"772b807b22ff5c57cc05cc803a8096552f75bd89"}]}