)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"a2a5e103a4a10399a3e8c4fd68732e86806d90ac","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Bump prlimit cpu time for qemu from 2 to 8"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"We\u0027ve got user reported bugs that when opperating with slow NFS"},{"line_number":10,"context_line":"backends with large (30+ GB) disk files, the prlimit of cpu_time 2 is"},{"line_number":11,"context_line":"not sufficient for completing the qemu-img command."},{"line_number":12,"context_line":""}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"3a71b18c_5fce0784","line":9,"range":{"start_line":9,"start_character":39,"end_line":9,"end_character":49},"updated":"2016-12-08 15:19:18.000000000","message":"operating","commit_id":"545f80e935d27e0dfcd013f05cc83e3496307373"}],"nova/virt/images.py":[{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"a2a5e103a4a10399a3e8c4fd68732e86806d90ac","unresolved":false,"context_lines":[{"line_number":39,"context_line":"IMAGE_API \u003d image.API()"},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"QEMU_IMG_LIMITS \u003d processutils.ProcessLimits("},{"line_number":42,"context_line":"    cpu_time\u003d8,"},{"line_number":43,"context_line":"    address_space\u003d1 * units.Gi)"},{"line_number":44,"context_line":""},{"line_number":45,"context_line":""}],"source_content_type":"text/x-python","patch_set":1,"id":"3a71b18c_1fdd6f42","line":42,"updated":"2016-12-08 15:19:18.000000000","message":"This seems to defeat the purpose of having low limits in the first place. Maybe we should make this configurable and if we fail with -9 then put something into the commit message saying you can bump the limit on this compute based on the config? That\u0027s not great though because you could just land the instance build on another compute that doesn\u0027t have this config, and then it\u0027s just a problem all over your deployment.\n\nCan we alternatively determine the cpu_time based on a size range of the image from the image meta? I think we\u0027re using qemu-img info to get the real size, but isn\u0027t there some base indication in the image meta?","commit_id":"545f80e935d27e0dfcd013f05cc83e3496307373"},{"author":{"_account_id":6962,"name":"Kashyap Chamarthy","email":"kchamart@redhat.com","username":"kashyapc"},"change_message_id":"5b3729bc3111f2ddb4f37a8d76ade4591b1237c6","unresolved":false,"context_lines":[{"line_number":39,"context_line":"IMAGE_API \u003d image.API()"},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"QEMU_IMG_LIMITS \u003d processutils.ProcessLimits("},{"line_number":42,"context_line":"    cpu_time\u003d8,"},{"line_number":43,"context_line":"    address_space\u003d1 * units.Gi)"},{"line_number":44,"context_line":""},{"line_number":45,"context_line":""}],"source_content_type":"text/x-python","patch_set":1,"id":"3a71b18c_128f1924","line":42,"in_reply_to":"3a71b18c_02851658","updated":"2016-12-08 19:26:04.000000000","message":"Yeah, I agree with the above.\n\n[The limit of \u00278\u0027 turned out to be an arbitrary guess I made, and incidentally, both the upstream \u0026 downstream reporters (different) said turning up the CPU limit resolved their failure scenarios.]","commit_id":"545f80e935d27e0dfcd013f05cc83e3496307373"},{"author":{"_account_id":9555,"name":"Matthew Booth","email":"mbooth@redhat.com","username":"MatthewBooth"},"change_message_id":"f671e2ead1980405a5a56901de015195c10fd12c","unresolved":false,"context_lines":[{"line_number":39,"context_line":"IMAGE_API \u003d image.API()"},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"QEMU_IMG_LIMITS \u003d processutils.ProcessLimits("},{"line_number":42,"context_line":"    cpu_time\u003d8,"},{"line_number":43,"context_line":"    address_space\u003d1 * units.Gi)"},{"line_number":44,"context_line":""},{"line_number":45,"context_line":""}],"source_content_type":"text/x-python","patch_set":1,"id":"3a71b18c_02851658","line":42,"in_reply_to":"3a71b18c_1fdd6f42","updated":"2016-12-08 15:37:08.000000000","message":"I don\u0027t think we want Yet Another Tuning Knob. It\u0027s ugly, but I think it\u0027s ok to increase this. The key thing is that there\u0027s a limit, which is worlds apart from no limit.","commit_id":"545f80e935d27e0dfcd013f05cc83e3496307373"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"a2a5e103a4a10399a3e8c4fd68732e86806d90ac","unresolved":false,"context_lines":[{"line_number":63,"context_line":"        out, err \u003d utils.execute(*cmd, prlimit\u003dQEMU_IMG_LIMITS)"},{"line_number":64,"context_line":"    except processutils.ProcessExecutionError as exp:"},{"line_number":65,"context_line":"        # this means we hit prlimits, make the exception more specific"},{"line_number":66,"context_line":"        if exp.exit_code \u003d\u003d -9:"},{"line_number":67,"context_line":"            msg \u003d (_(\"qemu-img aborted by prlimits when inspecting \""},{"line_number":68,"context_line":"                    \"%(path)s : %(exp)s\") % {\u0027path\u0027: path, \u0027exp\u0027: exp})"},{"line_number":69,"context_line":"        else:"}],"source_content_type":"text/x-python","patch_set":1,"id":"3a71b18c_bffe3bac","line":66,"updated":"2016-12-08 15:19:18.000000000","message":"We should have a unit test for this.","commit_id":"545f80e935d27e0dfcd013f05cc83e3496307373"}]}