)]}'
{"devstack-vm-gate.sh":[{"author":{"_account_id":2750,"name":"Sean Dague","email":"sean@dague.net","username":"sdague"},"change_message_id":"7a532db9e39ec2b3beb6e13b3c37db09ac32055f","unresolved":false,"context_lines":[{"line_number":647,"context_line":"            sed -ie \u0027s/ENABLED_SERVICES/SWIFT_STORAGE_SERVICES/g;s/DISABLED_SERVICES/ENABLED_SERVICES/g\u0027 localrc \\"},{"line_number":648,"context_line":"            executable\u003d/bin/bash\" \\"},{"line_number":649,"context_line":"            \u0026\u003e\u003e \"$WORKSPACE/logs/devstack-subnodes-early-swift.txt\""},{"line_number":650,"context_line":"    fi"},{"line_number":651,"context_line":"    $ANSIBLE primary -f 5 -i \"$WORKSPACE/inventory\" -m shell \\"},{"line_number":652,"context_line":"        -a \"cd \u0027$BASE/new/devstack\u0027 \u0026\u0026 sudo -H -u stack stdbuf -oL -eL ./stack.sh executable\u003d/bin/bash\" \\"},{"line_number":653,"context_line":"        \u0026\u003e \"$WORKSPACE/logs/devstack-early.txt\""}],"source_content_type":"text/x-sh","patch_set":13,"id":"5a18252c_b5126507","line":650,"updated":"2016-04-14 12:02:44.000000000","message":"Instead of this would it be possible to trigger a ring rebuild after the subnode deploys? (even if that means we\u0027ve built the ring twice?)\n\nBecause the node build order today is:\n\n - build primary completely\n - build subnodes one at a time\n\nAnd changing that order is going to make other things complicated.","commit_id":"5eb8642b765f7724e172338df14d92f9559d1f83"},{"author":{"_account_id":6968,"name":"Christian Schwede","email":"cschwede@nvidia.com","username":"cschwede"},"change_message_id":"c6702a2f2977d10263bebd6f91a52a122cad9a0b","unresolved":false,"context_lines":[{"line_number":647,"context_line":"            sed -ie \u0027s/ENABLED_SERVICES/SWIFT_STORAGE_SERVICES/g;s/DISABLED_SERVICES/ENABLED_SERVICES/g\u0027 localrc \\"},{"line_number":648,"context_line":"            executable\u003d/bin/bash\" \\"},{"line_number":649,"context_line":"            \u0026\u003e\u003e \"$WORKSPACE/logs/devstack-subnodes-early-swift.txt\""},{"line_number":650,"context_line":"    fi"},{"line_number":651,"context_line":"    $ANSIBLE primary -f 5 -i \"$WORKSPACE/inventory\" -m shell \\"},{"line_number":652,"context_line":"        -a \"cd \u0027$BASE/new/devstack\u0027 \u0026\u0026 sudo -H -u stack stdbuf -oL -eL ./stack.sh executable\u003d/bin/bash\" \\"},{"line_number":653,"context_line":"        \u0026\u003e \"$WORKSPACE/logs/devstack-early.txt\""}],"source_content_type":"text/x-sh","patch_set":13,"id":"5a18252c_2b54321f","line":650,"in_reply_to":"5a18252c_b5126507","updated":"2016-04-14 12:39:12.000000000","message":"Well, not without some possible side effects.\n\nWhen the primary node gets deployed, some data will be alredy stored in Swift (an image upload). So if we change the ring later on we need to wait until replicators moved all data from the primary node to the subnode, otherwise it will become unavailable. However, I\u0027m not sure if this is used later on in Tempest? We could give it a try.\n\nWhat we could do is to include both the primary and subnode in the ring, and later stop the storage services on the primary.","commit_id":"5eb8642b765f7724e172338df14d92f9559d1f83"}]}