)]}'
{"specs/mitaka/volume-fail-reason.rst":[{"author":{"_account_id":4523,"name":"Eric Harney","email":"eharney@redhat.com","username":"eharney"},"change_message_id":"ad64b0a48acfb8152fe0f3e4556ec0270204021f","unresolved":false,"context_lines":[{"line_number":30,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":31,"context_line":""},{"line_number":32,"context_line":"After each error with volume modification we should store error message in the"},{"line_number":33,"context_line":"`fail_reason` column of the Volume and Snapshot tables."},{"line_number":34,"context_line":""},{"line_number":35,"context_line":""},{"line_number":36,"context_line":"Alternatives"}],"source_content_type":"text/x-rst","patch_set":4,"id":"3a29b11f_a0007302","line":33,"updated":"2015-10-20 14:36:34.000000000","message":"I don\u0027t think this would work so well for volumes and snapshots.  There are many different operations that can fail, and that can be performed simultaneously or in quick succession.  This scheme may work most of the time for backups because there aren\u0027t that many operations, but it looks somewhat iffy even there.\n\nI think that, at scale, this will just end up being rather confusing, because if you run 5 operations over a short period of time, and two of them fail, which operation does this error relate to?  You can only tell if you ensure a request id is stored with the message.\n\nThe proposal we discussed a couple of summits ago seemed like a better fit for what we need:\n\nhttps://etherpad.openstack.org/p/kilo-cinder-async-reporting\n\nAlso related:\n\nhttps://etherpad.openstack.org/p/liberty-cross-project-user-notifications","commit_id":"1c14df9cec28eb5d850be6bea76959ba044292be"},{"author":{"_account_id":4523,"name":"Eric Harney","email":"eharney@redhat.com","username":"eharney"},"change_message_id":"ad64b0a48acfb8152fe0f3e4556ec0270204021f","unresolved":false,"context_lines":[{"line_number":35,"context_line":""},{"line_number":36,"context_line":"Alternatives"},{"line_number":37,"context_line":"------------"},{"line_number":38,"context_line":""},{"line_number":39,"context_line":"* Leave as is and store such info only in logs"},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"* Implement some async error reporting mechanism"}],"source_content_type":"text/x-rst","patch_set":4,"id":"3a29b11f_e0baeb22","line":38,"updated":"2015-10-20 14:36:34.000000000","message":"Another alternative (discussed previously when working on this same area): use something based on our notifications system.  (I don\u0027t think this is the way to go, but it was part of this discussion.)","commit_id":"1c14df9cec28eb5d850be6bea76959ba044292be"},{"author":{"_account_id":4523,"name":"Eric Harney","email":"eharney@redhat.com","username":"eharney"},"change_message_id":"ad64b0a48acfb8152fe0f3e4556ec0270204021f","unresolved":false,"context_lines":[{"line_number":38,"context_line":""},{"line_number":39,"context_line":"* Leave as is and store such info only in logs"},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"* Implement some async error reporting mechanism"},{"line_number":42,"context_line":""},{"line_number":43,"context_line":"Data model impact"},{"line_number":44,"context_line":"-----------------"}],"source_content_type":"text/x-rst","patch_set":4,"id":"3a29b11f_7d2a042e","line":41,"updated":"2015-10-20 14:36:34.000000000","message":"I suspect that this is what we actually need -- a mechanism that tracks errors per-operation, not per-volume.","commit_id":"1c14df9cec28eb5d850be6bea76959ba044292be"}]}
