)]}'
{"mariadb/templates/bin/_start.py.tpl":[{"author":{"_account_id":8898,"name":"Chris Wedgwood","email":"cw@f00f.org","username":"anticw"},"change_message_id":"3fc84c907bc0e7a4737f4fe12e0e7f3cfdb81b96","unresolved":false,"context_lines":[{"line_number":329,"context_line":"            logger.warning(\"Collision writing configmap: {0}\".format(error))"},{"line_number":330,"context_line":"            # This often happens when the replicas were started at the same"},{"line_number":331,"context_line":"            # time, and tends to be persistent. Sleep briefly to break the"},{"line_number":332,"context_line":"            # synchronization."},{"line_number":333,"context_line":"            time.sleep(1)"},{"line_number":334,"context_line":"            return True"},{"line_number":335,"context_line":"        else:"},{"line_number":336,"context_line":"            logger.error(\"Failed to set configmap: {0}\".format(error))"}],"source_content_type":"text/x-smarty","patch_set":1,"id":"1fa4df85_b168fc1c","line":333,"range":{"start_line":332,"start_character":13,"end_line":333,"end_character":25},"updated":"2020-03-04 19:55:10.000000000","message":"why not a random delay on startup so they end up staggered?","commit_id":"2a4b369f25f2c2cf3880f9c4a1a9ca0d5e0b6eb8"},{"author":{"_account_id":28719,"name":"Phil Sphicas","email":"phil.sphicas@att.com","username":"ps3910"},"change_message_id":"807ca850f2bacae97c902d14c72057456dcbf561","unresolved":false,"context_lines":[{"line_number":329,"context_line":"            logger.warning(\"Collision writing configmap: {0}\".format(error))"},{"line_number":330,"context_line":"            # This often happens when the replicas were started at the same"},{"line_number":331,"context_line":"            # time, and tends to be persistent. Sleep briefly to break the"},{"line_number":332,"context_line":"            # synchronization."},{"line_number":333,"context_line":"            time.sleep(1)"},{"line_number":334,"context_line":"            return True"},{"line_number":335,"context_line":"        else:"},{"line_number":336,"context_line":"            logger.error(\"Failed to set configmap: {0}\".format(error))"}],"source_content_type":"text/x-smarty","patch_set":1,"id":"1fa4df85_63a2b9b0","line":333,"range":{"start_line":332,"start_character":13,"end_line":333,"end_character":25},"in_reply_to":"1fa4df85_b168fc1c","updated":"2020-03-04 22:42:34.000000000","message":"That could make the issue less likely to occur initially if the pods are all starting at the same time, but if we get unlucky (e.g. after killing a pod), then they will stay in sync. We would need the attempts offset by at least the duration of a successful end-to-end update of the grastate configmap (involving a separate read and write). If that is ~10ms, and we have three replicas, we would still have a ~0.3% chance of hitting the issue (any two or more replicas landing in the same 10ms slot). Worse if the update takes longer.","commit_id":"2a4b369f25f2c2cf3880f9c4a1a9ca0d5e0b6eb8"}]}