)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"b7f85551f2edb8526b812f7ca3d407165fe5f93c","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Set defult value of num_nvme_discover_tries\u003d5"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Discovering newly connected devices in the initiator side can be slow"},{"line_number":10,"context_line":"when there are old connections, increasing the retries is important "},{"line_number":11,"context_line":"to discover the new connected device."},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Change-Id: I62a8162bf96d51f7252cfefbffc1a46010a3a612"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":2,"id":"3f79a3b5_e6e54aa5","line":10,"range":{"start_line":9,"start_character":0,"end_line":10,"end_character":30},"updated":"2018-09-19 15:52:33.000000000","message":"Do we have a predictable failure when this happens? For example, you\u0027re raising the default from 3 to 5 for retries, but what happens when I just pile up more old connections on my compute? It\u0027s still going to possibly hit the same issue, right? In other words, are we going to see another patch in a year from now to raise this to 10 retries by default or something? What I\u0027d like to see, if it\u0027s possible, would be some code which traps a known failure when we timeout and then logs a suggestion for the operator reading the error that if this is a persistent problem, they should consider increasing the num_nvme_discover_tries value. Then it\u0027s a bit more serviceable for the operator rather than us having to guess at what the best default is.\n\nAs for the release note, I don\u0027t really care about that for this simple change - I mostly care that we will continue to see this and operators will have to hunt through the code to figure out what to modify to ease the pain.","commit_id":"3574f8970a5647efa0c28ab68dc2672ab3eccb87"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"b7f85551f2edb8526b812f7ca3d407165fe5f93c","unresolved":false,"context_lines":[{"line_number":7,"context_line":"Set defult value of num_nvme_discover_tries\u003d5"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Discovering newly connected devices in the initiator side can be slow"},{"line_number":10,"context_line":"when there are old connections, increasing the retries is important "},{"line_number":11,"context_line":"to discover the new connected device."},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Change-Id: I62a8162bf96d51f7252cfefbffc1a46010a3a612"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":2,"id":"3f79a3b5_e6578a59","line":10,"range":{"start_line":10,"start_character":67,"end_line":10,"end_character":68},"updated":"2018-09-19 15:52:33.000000000","message":"whitespace damage","commit_id":"3574f8970a5647efa0c28ab68dc2672ab3eccb87"},{"author":{"_account_id":22948,"name":"Hamdy Khader","email":"hamdyk@mellanox.com","username":"hamdyk"},"change_message_id":"23d0d3a4edf32547a267fed9047d49e3c9151b0c","unresolved":false,"context_lines":[{"line_number":7,"context_line":"Set defult value of num_nvme_discover_tries\u003d5"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Discovering newly connected devices in the initiator side can be slow"},{"line_number":10,"context_line":"when there are old connections, increasing the retries is important "},{"line_number":11,"context_line":"to discover the new connected device."},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Change-Id: I62a8162bf96d51f7252cfefbffc1a46010a3a612"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":2,"id":"3f79a3b5_f1713fe1","line":10,"range":{"start_line":10,"start_character":67,"end_line":10,"end_character":68},"in_reply_to":"3f79a3b5_e6578a59","updated":"2018-09-23 10:52:33.000000000","message":"Done","commit_id":"3574f8970a5647efa0c28ab68dc2672ab3eccb87"},{"author":{"_account_id":22948,"name":"Hamdy Khader","email":"hamdyk@mellanox.com","username":"hamdyk"},"change_message_id":"23d0d3a4edf32547a267fed9047d49e3c9151b0c","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Set defult value of num_nvme_discover_tries\u003d5"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Discovering newly connected devices in the initiator side can be slow"},{"line_number":10,"context_line":"when there are old connections, increasing the retries is important "},{"line_number":11,"context_line":"to discover the new connected device."},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Change-Id: I62a8162bf96d51f7252cfefbffc1a46010a3a612"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":2,"id":"3f79a3b5_5146f37d","line":10,"range":{"start_line":9,"start_character":0,"end_line":10,"end_character":30},"in_reply_to":"3f79a3b5_e6e54aa5","updated":"2018-09-23 10:52:33.000000000","message":"We ran some tests on different environments, the longest waiting time was 28 seconds.\nThat means the default retry count of 5 will give us 60 seconds of sleep time which is more than enough for nvme to restore all connections.\nI like your idea about suggesting to the operator to increase the count, will do in os-brick change:\nhttps://review.openstack.org/#/c/602332/","commit_id":"3574f8970a5647efa0c28ab68dc2672ab3eccb87"}]}