)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"2ff23ff381de5fa9e9c13f97844ef675d677fed4","unresolved":false,"context_lines":[{"line_number":9,"context_line":"This implements an auto-disable feature in nova-compute, where we"},{"line_number":10,"context_line":"automatically set our service record to disabled if we consecutively"},{"line_number":11,"context_line":"fail to build a certain number of instances."},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Change-Id: I02b7cd87d399d487dd1d650540f503a70bc27749"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":3,"id":"3f044301_93262bce","line":12,"updated":"2017-05-10 11:18:18.000000000","message":"I think it would be fair to link this to the claims in the scheduler blueprint. It really came up in the cells v2 discussion at the summit, but it was related to talking about claims in the scheduler.","commit_id":"38fb5f3a012d7fd62a0d97ef74826b5868f49ca5"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"ba438041e9a626c285ed7b24541ed73bdd1161b8","unresolved":false,"context_lines":[{"line_number":9,"context_line":"This implements an auto-disable feature in nova-compute, where we"},{"line_number":10,"context_line":"automatically set our service record to disabled if we consecutively"},{"line_number":11,"context_line":"fail to build a certain number of instances."},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Change-Id: I02b7cd87d399d487dd1d650540f503a70bc27749"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":3,"id":"3f044301_8996fe4b","line":12,"in_reply_to":"3f044301_93262bce","updated":"2017-05-10 12:28:50.000000000","message":"Really? I don\u0027t recall this having much to do with that, other than that claims will reduce the need for reschedules at all down to very few.","commit_id":"38fb5f3a012d7fd62a0d97ef74826b5868f49ca5"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"f1272a874749169ac9ecd272210a153ae8e3a1bf","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Make compute auto-disable itself if builds are failing"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"This implements an auto-disable feature in nova-compute, where we"},{"line_number":10,"context_line":"automatically set our service record to disabled if we consecutively"},{"line_number":11,"context_line":"fail to build a certain number of instances."},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Change-Id: I02b7cd87d399d487dd1d650540f503a70bc27749"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":4,"id":"1f013ff3_0fd035f7","line":11,"range":{"start_line":9,"start_character":0,"end_line":11,"end_character":44},"updated":"2017-05-12 17:54:13.000000000","message":"nit: this seems fairly random to anyone that wasn\u0027t at the summit, and I think it would be good to note the background discussion here where operators were talking about how an almost full compute node becomes a magnet for collisions in a pack-first scheduling scheme. That\u0027s really what prompted this (in addition to talking about removing retries for cells v2 because of the up-call, and claims in the scheduler).","commit_id":"169576f45124d9300c81b0a6087d5c76ba30a342"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"46cdc1a05db26d62f96fc9ebd04ff6fe7230a1f8","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Make compute auto-disable itself if builds are failing"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"This implements an auto-disable feature in nova-compute, where we"},{"line_number":10,"context_line":"automatically set our service record to disabled if we consecutively"},{"line_number":11,"context_line":"fail to build a certain number of instances."},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Change-Id: I02b7cd87d399d487dd1d650540f503a70bc27749"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":5,"id":"1f013ff3_6f7d79e6","line":11,"range":{"start_line":9,"start_character":0,"end_line":11,"end_character":44},"updated":"2017-05-12 17:58:40.000000000","message":"Repeating from PS4:\n\nThis seems fairly random to anyone that wasn\u0027t at the summit, and I think it would be good to note the background discussion here where operators were talking about how an almost full compute node becomes a magnet for collisions in a pack-first scheduling scheme. That\u0027s really what prompted this (in addition to talking about removing retries for cells v2 because of the up-call, and claims in the scheduler).","commit_id":"73c180d6ee8b4ac8517ed44e1d74c1613a90bb02"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"126aac24211ff5a452292f4e0e0f50d0ce319ad0","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Make compute auto-disable itself if builds are failing"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"This implements an auto-disable feature in nova-compute, where we"},{"line_number":10,"context_line":"automatically set our service record to disabled if we consecutively"},{"line_number":11,"context_line":"fail to build a certain number of instances."},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Change-Id: I02b7cd87d399d487dd1d650540f503a70bc27749"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":5,"id":"1f013ff3_2e7633a6","line":11,"range":{"start_line":9,"start_character":0,"end_line":11,"end_character":44},"in_reply_to":"1f013ff3_6f7d79e6","updated":"2017-05-15 16:17:10.000000000","message":"I think it was something they all wanted to see independent of the retries discussion, FWIW. I can add more words about this being more important in the future, but I would expect everyone would turn/leave this on immediately if it were available because it just makes a lot of sense.","commit_id":"73c180d6ee8b4ac8517ed44e1d74c1613a90bb02"}],"nova/compute/manager.py":[{"author":{"_account_id":8768,"name":"Chris Friesen","email":"chris.friesen@windriver.com","username":"cbf123"},"change_message_id":"3b2f59cbd562830d408e4a9527491ea646fd2cba","unresolved":false,"context_lines":[{"line_number":529,"context_line":"        else:"},{"line_number":530,"context_line":"            self._live_migration_semaphore \u003d compute_utils.UnlimitedSemaphore()"},{"line_number":531,"context_line":""},{"line_number":532,"context_line":"        self._failed_builds \u003d 0"},{"line_number":533,"context_line":""},{"line_number":534,"context_line":"        super(ComputeManager, self).__init__(service_name\u003d\"compute\","},{"line_number":535,"context_line":"                                             *args, **kwargs)"}],"source_content_type":"text/x-python","patch_set":2,"id":"3f044301_856a6fb0","line":532,"updated":"2017-05-09 18:04:37.000000000","message":"nit: remove blank line before this?","commit_id":"2bf79e899c2e8a320ffbae7154fcdee613eca8d8"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"4d1f6bf9e0f1fbfbd45759f805cb5f9c8d1f96e6","unresolved":false,"context_lines":[{"line_number":529,"context_line":"        else:"},{"line_number":530,"context_line":"            self._live_migration_semaphore \u003d compute_utils.UnlimitedSemaphore()"},{"line_number":531,"context_line":""},{"line_number":532,"context_line":"        self._failed_builds \u003d 0"},{"line_number":533,"context_line":""},{"line_number":534,"context_line":"        super(ComputeManager, self).__init__(service_name\u003d\"compute\","},{"line_number":535,"context_line":"                                             *args, **kwargs)"}],"source_content_type":"text/x-python","patch_set":2,"id":"3f044301_00776da2","line":532,"in_reply_to":"3f044301_856a6fb0","updated":"2017-05-09 18:39:17.000000000","message":"Meh.","commit_id":"2bf79e899c2e8a320ffbae7154fcdee613eca8d8"},{"author":{"_account_id":3189,"name":"Chet Burgess","email":"cfb@metacloud.com","username":"cfb-n"},"change_message_id":"40824b0954a939e87dc0be8f0e6f6537909228cc","unresolved":false,"context_lines":[{"line_number":1690,"context_line":"        return block_device_info"},{"line_number":1691,"context_line":""},{"line_number":1692,"context_line":"    def _build_failed(self):"},{"line_number":1693,"context_line":"        self._failed_builds +\u003d 1"},{"line_number":1694,"context_line":"        limit \u003d CONF.consecutive_build_service_disable_threshold"},{"line_number":1695,"context_line":"        if limit and self._failed_builds \u003e\u003d limit:"},{"line_number":1696,"context_line":"            LOG.error(\u0027Disabling service due to %(fails)i \u0027"}],"source_content_type":"text/x-python","patch_set":4,"id":"3f044301_87c7fe64","line":1693,"range":{"start_line":1693,"start_character":0,"end_line":1693,"end_character":32},"updated":"2017-05-10 18:24:02.000000000","message":"Can we have a simple log line indicating that the count is incrementing and what the new count is?","commit_id":"169576f45124d9300c81b0a6087d5c76ba30a342"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"0c0460c0c4b9d3645169d88e3542b228a8a44b67","unresolved":false,"context_lines":[{"line_number":1690,"context_line":"        return block_device_info"},{"line_number":1691,"context_line":""},{"line_number":1692,"context_line":"    def _build_failed(self):"},{"line_number":1693,"context_line":"        self._failed_builds +\u003d 1"},{"line_number":1694,"context_line":"        limit \u003d CONF.consecutive_build_service_disable_threshold"},{"line_number":1695,"context_line":"        if limit and self._failed_builds \u003e\u003d limit:"},{"line_number":1696,"context_line":"            LOG.error(\u0027Disabling service due to %(fails)i \u0027"}],"source_content_type":"text/x-python","patch_set":4,"id":"3f044301_ad040619","line":1693,"range":{"start_line":1693,"start_character":0,"end_line":1693,"end_character":32},"in_reply_to":"3f044301_87c7fe64","updated":"2017-05-11 19:31:42.000000000","message":"Like the one on L1709?","commit_id":"169576f45124d9300c81b0a6087d5c76ba30a342"},{"author":{"_account_id":3189,"name":"Chet Burgess","email":"cfb@metacloud.com","username":"cfb-n"},"change_message_id":"40824b0954a939e87dc0be8f0e6f6537909228cc","unresolved":false,"context_lines":[{"line_number":1692,"context_line":"    def _build_failed(self):"},{"line_number":1693,"context_line":"        self._failed_builds +\u003d 1"},{"line_number":1694,"context_line":"        limit \u003d CONF.consecutive_build_service_disable_threshold"},{"line_number":1695,"context_line":"        if limit and self._failed_builds \u003e\u003d limit:"},{"line_number":1696,"context_line":"            LOG.error(\u0027Disabling service due to %(fails)i \u0027"},{"line_number":1697,"context_line":"                      \u0027consecuitive build failures\u0027,"},{"line_number":1698,"context_line":"                      {\u0027fails\u0027: self._failed_builds})"}],"source_content_type":"text/x-python","patch_set":4,"id":"3f044301_8775be34","line":1695,"range":{"start_line":1695,"start_character":0,"end_line":1695,"end_character":50},"updated":"2017-05-10 18:24:02.000000000","message":"I think there is a corner-case here where a build could fail *after* the threshold has been hit that would increment the counter again.\n\nGiven that the next successful boot will clear the count this is probably fine, but it might be worth a simple comment indicating that its possible but not an issue due to the clear-on-success later.","commit_id":"169576f45124d9300c81b0a6087d5c76ba30a342"},{"author":{"_account_id":23882,"name":"Dan Peschman","email":"dpeschman@godaddy.com","username":"dtp"},"change_message_id":"f293e4d007a5ffd4754ea17fc08589c837913aa4","unresolved":false,"context_lines":[{"line_number":1694,"context_line":"        limit \u003d CONF.consecutive_build_service_disable_threshold"},{"line_number":1695,"context_line":"        if limit and self._failed_builds \u003e\u003d limit:"},{"line_number":1696,"context_line":"            LOG.error(\u0027Disabling service due to %(fails)i \u0027"},{"line_number":1697,"context_line":"                      \u0027consecuitive build failures\u0027,"},{"line_number":1698,"context_line":"                      {\u0027fails\u0027: self._failed_builds})"},{"line_number":1699,"context_line":"            ctx \u003d nova.context.get_admin_context()"},{"line_number":1700,"context_line":"            service \u003d objects.Service.get_by_compute_host(ctx, CONF.host)"}],"source_content_type":"text/x-python","patch_set":4,"id":"3f044301_8726e514","line":1697,"updated":"2017-05-10 13:50:47.000000000","message":"\"consecutive\"","commit_id":"169576f45124d9300c81b0a6087d5c76ba30a342"},{"author":{"_account_id":3189,"name":"Chet Burgess","email":"cfb@metacloud.com","username":"cfb-n"},"change_message_id":"40824b0954a939e87dc0be8f0e6f6537909228cc","unresolved":false,"context_lines":[{"line_number":1732,"context_line":"                    raise"},{"line_number":1733,"context_line":"                finally:"},{"line_number":1734,"context_line":"                    fails \u003d (build_results.FAILED,"},{"line_number":1735,"context_line":"                             build_results.RESCHEDULED)"},{"line_number":1736,"context_line":"                    if result in fails:"},{"line_number":1737,"context_line":"                        self._build_failed()"},{"line_number":1738,"context_line":"                    else:"}],"source_content_type":"text/x-python","patch_set":4,"id":"3f044301_07588ef7","line":1735,"range":{"start_line":1735,"start_character":0,"end_line":1735,"end_character":55},"updated":"2017-05-10 18:24:02.000000000","message":"Do we want to include RESCHEDULED at this time given we have known corner case issues with the scheduler sub-optimally picking nodes under full-ish conditions?","commit_id":"169576f45124d9300c81b0a6087d5c76ba30a342"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"0c0460c0c4b9d3645169d88e3542b228a8a44b67","unresolved":false,"context_lines":[{"line_number":1732,"context_line":"                    raise"},{"line_number":1733,"context_line":"                finally:"},{"line_number":1734,"context_line":"                    fails \u003d (build_results.FAILED,"},{"line_number":1735,"context_line":"                             build_results.RESCHEDULED)"},{"line_number":1736,"context_line":"                    if result in fails:"},{"line_number":1737,"context_line":"                        self._build_failed()"},{"line_number":1738,"context_line":"                    else:"}],"source_content_type":"text/x-python","patch_set":4,"id":"3f044301_306de9cb","line":1735,"range":{"start_line":1735,"start_character":0,"end_line":1735,"end_character":55},"in_reply_to":"3f044301_07588ef7","updated":"2017-05-11 19:31:42.000000000","message":"I\u0027m tempted to leave this in since we\u0027re hoping to merge scheduler claims in the next month or so. Also, if you have 10 consecutive reschedules for any reason, things might be broken in some resource-reporting sense too.","commit_id":"169576f45124d9300c81b0a6087d5c76ba30a342"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"f1272a874749169ac9ecd272210a153ae8e3a1bf","unresolved":false,"context_lines":[{"line_number":1732,"context_line":"                    raise"},{"line_number":1733,"context_line":"                finally:"},{"line_number":1734,"context_line":"                    fails \u003d (build_results.FAILED,"},{"line_number":1735,"context_line":"                             build_results.RESCHEDULED)"},{"line_number":1736,"context_line":"                    if result in fails:"},{"line_number":1737,"context_line":"                        self._build_failed()"},{"line_number":1738,"context_line":"                    else:"}],"source_content_type":"text/x-python","patch_set":4,"id":"1f013ff3_0fcb9553","line":1735,"range":{"start_line":1735,"start_character":0,"end_line":1735,"end_character":55},"in_reply_to":"3f044301_306de9cb","updated":"2017-05-12 17:54:13.000000000","message":"This was bothering me too since we don\u0027t have claims in the scheduler yet. I know we\u0027re targeting that for Pike, but it\u0027s not in yet, so I don\u0027t want to write code today assuming that\u0027s going to be in anytime soon. That\u0027s why I originally thought we should ship this with it disabled by default so it doesn\u0027t catch people off guard.\n\nHowever, I had to remember that the other reason this is being added is specifically for the packing / collision case that was pointed out in the cells v2 session at the Boston Forum, which was where this compute node is mostly full and the scheduler keeps picking it for builds which just makes this node a magnet for failures and it\u0027d be better to just disable it if that\u0027s the case. If operators don\u0027t like the auto-disable they can disable this option or bump it way up to like 50 or something as a last resort.\n\nSo having said that, I think I\u0027m OK with this now. It\u0027s just important to keep that scenario/context in mind.","commit_id":"169576f45124d9300c81b0a6087d5c76ba30a342"},{"author":{"_account_id":14070,"name":"Eric Fried","email":"openstack@fried.cc","username":"efried"},"change_message_id":"1a4de212390734079077cf882afd6657eec1b3ed","unresolved":false,"context_lines":[{"line_number":1736,"context_line":"                    result \u003d build_results.FAILED"},{"line_number":1737,"context_line":"                    raise"},{"line_number":1738,"context_line":"                finally:"},{"line_number":1739,"context_line":"                    fails \u003d (build_results.FAILED,"},{"line_number":1740,"context_line":"                             build_results.RESCHEDULED)"},{"line_number":1741,"context_line":"                    if result in fails:"},{"line_number":1742,"context_line":"                        self._build_failed()"},{"line_number":1743,"context_line":"                    else:"}],"source_content_type":"text/x-python","patch_set":6,"id":"ff0f0b1f_05f9e6c6","line":1740,"range":{"start_line":1739,"start_character":20,"end_line":1740,"end_character":55},"updated":"2017-05-19 15:53:45.000000000","message":"Would rather see this in a const.","commit_id":"f93f675a8563c7c37fdcb0c685a8b491a7311361"}],"nova/conf/compute.py":[{"author":{"_account_id":8768,"name":"Chris Friesen","email":"chris.friesen@windriver.com","username":"cbf123"},"change_message_id":"3b2f59cbd562830d408e4a9527491ea646fd2cba","unresolved":false,"context_lines":[{"line_number":641,"context_line":"Number of consecutive failed builds that result in disabing a compute service."},{"line_number":642,"context_line":""},{"line_number":643,"context_line":"This option will cause nova-compute to set itself to a disabled state"},{"line_number":644,"context_line":"if a certain number of consecutive build faulures occur. This will"},{"line_number":645,"context_line":"the scheduler from continuing to send builds to a compute node that is"},{"line_number":646,"context_line":"consistently failing."},{"line_number":647,"context_line":""}],"source_content_type":"text/x-python","patch_set":2,"id":"3f044301_45e7c741","line":644,"range":{"start_line":644,"start_character":62,"end_line":644,"end_character":66},"updated":"2017-05-09 18:04:37.000000000","message":"should be something like \"This will prevent the scheduler...\"","commit_id":"2bf79e899c2e8a320ffbae7154fcdee613eca8d8"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"4d1f6bf9e0f1fbfbd45759f805cb5f9c8d1f96e6","unresolved":false,"context_lines":[{"line_number":641,"context_line":"Number of consecutive failed builds that result in disabing a compute service."},{"line_number":642,"context_line":""},{"line_number":643,"context_line":"This option will cause nova-compute to set itself to a disabled state"},{"line_number":644,"context_line":"if a certain number of consecutive build faulures occur. This will"},{"line_number":645,"context_line":"the scheduler from continuing to send builds to a compute node that is"},{"line_number":646,"context_line":"consistently failing."},{"line_number":647,"context_line":""}],"source_content_type":"text/x-python","patch_set":2,"id":"3f044301_c0574502","line":644,"range":{"start_line":644,"start_character":62,"end_line":644,"end_character":66},"in_reply_to":"3f044301_45e7c741","updated":"2017-05-09 18:39:17.000000000","message":"Yep, typing-at-summit brain.","commit_id":"2bf79e899c2e8a320ffbae7154fcdee613eca8d8"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"2ff23ff381de5fa9e9c13f97844ef675d677fed4","unresolved":false,"context_lines":[{"line_number":635,"context_line":""},{"line_number":636,"context_line":"* Any positive integer representing greenthreads count."},{"line_number":637,"context_line":"\"\"\"),"},{"line_number":638,"context_line":"    cfg.IntOpt(\u0027consecutive_build_service_disable_threshold\u0027,"},{"line_number":639,"context_line":"        default\u003d10,"},{"line_number":640,"context_line":"        help\u003d\"\"\""},{"line_number":641,"context_line":"Number of consecutive failed builds that result in disabing a compute service."}],"source_content_type":"text/x-python","patch_set":3,"id":"3f044301_134e9b96","line":638,"range":{"start_line":638,"start_character":16,"end_line":638,"end_character":59},"updated":"2017-05-10 11:18:18.000000000","message":"nit: I\u0027d ask sfinucan, but I think at some point we want to put these compute-specific config options into a [compute] group, so maybe we should start that here.","commit_id":"38fb5f3a012d7fd62a0d97ef74826b5868f49ca5"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"3e7703094c17593164dd60066bb0af0bc5a8ffaf","unresolved":false,"context_lines":[{"line_number":635,"context_line":""},{"line_number":636,"context_line":"* Any positive integer representing greenthreads count."},{"line_number":637,"context_line":"\"\"\"),"},{"line_number":638,"context_line":"    cfg.IntOpt(\u0027consecutive_build_service_disable_threshold\u0027,"},{"line_number":639,"context_line":"        default\u003d10,"},{"line_number":640,"context_line":"        help\u003d\"\"\""},{"line_number":641,"context_line":"Number of consecutive failed builds that result in disabing a compute service."}],"source_content_type":"text/x-python","patch_set":3,"id":"3f044301_e2ceb72b","line":638,"range":{"start_line":638,"start_character":16,"end_line":638,"end_character":59},"in_reply_to":"3f044301_134e9b96","updated":"2017-05-10 14:34:53.000000000","message":"Yup, let\u0027s start here. Look at nova/conf/libvirt.py for an example of how to declare an OptGroup and how to register option(s) to that group. You\u0027ll need to update any references to it from \u0027CONF.consecutive_build_service_disable_threshold\u0027 to \u0027CONF.compute.consecutive_build_service_disable_threshold\u0027","commit_id":"38fb5f3a012d7fd62a0d97ef74826b5868f49ca5"},{"author":{"_account_id":23882,"name":"Dan Peschman","email":"dpeschman@godaddy.com","username":"dtp"},"change_message_id":"e460682cbadb9054af9e0162f3ae1c2793d5f6a5","unresolved":false,"context_lines":[{"line_number":636,"context_line":"* Any positive integer representing greenthreads count."},{"line_number":637,"context_line":"\"\"\"),"},{"line_number":638,"context_line":"    cfg.IntOpt(\u0027consecutive_build_service_disable_threshold\u0027,"},{"line_number":639,"context_line":"        default\u003d10,"},{"line_number":640,"context_line":"        help\u003d\"\"\""},{"line_number":641,"context_line":"Number of consecutive failed builds that result in disabing a compute service."},{"line_number":642,"context_line":""}],"source_content_type":"text/x-python","patch_set":3,"id":"3f044301_4ff8017a","line":639,"updated":"2017-05-10 03:21:48.000000000","message":"default 0 for backwards compat?  (or does it simply work that way if they don\u0027t have the config entry?)","commit_id":"38fb5f3a012d7fd62a0d97ef74826b5868f49ca5"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"ba438041e9a626c285ed7b24541ed73bdd1161b8","unresolved":false,"context_lines":[{"line_number":636,"context_line":"* Any positive integer representing greenthreads count."},{"line_number":637,"context_line":"\"\"\"),"},{"line_number":638,"context_line":"    cfg.IntOpt(\u0027consecutive_build_service_disable_threshold\u0027,"},{"line_number":639,"context_line":"        default\u003d10,"},{"line_number":640,"context_line":"        help\u003d\"\"\""},{"line_number":641,"context_line":"Number of consecutive failed builds that result in disabing a compute service."},{"line_number":642,"context_line":""}],"source_content_type":"text/x-python","patch_set":3,"id":"3f044301_499c7669","line":639,"in_reply_to":"3f044301_13f21b71","updated":"2017-05-10 12:28:50.000000000","message":"I think this always makes sense so I\u0027d rather have a sane default, so I\u0027ll add it to the reno.","commit_id":"38fb5f3a012d7fd62a0d97ef74826b5868f49ca5"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"2ff23ff381de5fa9e9c13f97844ef675d677fed4","unresolved":false,"context_lines":[{"line_number":636,"context_line":"* Any positive integer representing greenthreads count."},{"line_number":637,"context_line":"\"\"\"),"},{"line_number":638,"context_line":"    cfg.IntOpt(\u0027consecutive_build_service_disable_threshold\u0027,"},{"line_number":639,"context_line":"        default\u003d10,"},{"line_number":640,"context_line":"        help\u003d\"\"\""},{"line_number":641,"context_line":"Number of consecutive failed builds that result in disabing a compute service."},{"line_number":642,"context_line":""}],"source_content_type":"text/x-python","patch_set":3,"id":"3f044301_13f21b71","line":639,"in_reply_to":"3f044301_4ff8017a","updated":"2017-05-10 11:18:18.000000000","message":"They\u0027ll have the option by default, so yeah either this needs to be 0 for backward compat, or we need to also call this out as a new default in the \u0027upgrades\u0027 section of the release note.","commit_id":"38fb5f3a012d7fd62a0d97ef74826b5868f49ca5"},{"author":{"_account_id":23882,"name":"Dan Peschman","email":"dpeschman@godaddy.com","username":"dtp"},"change_message_id":"e460682cbadb9054af9e0162f3ae1c2793d5f6a5","unresolved":false,"context_lines":[{"line_number":638,"context_line":"    cfg.IntOpt(\u0027consecutive_build_service_disable_threshold\u0027,"},{"line_number":639,"context_line":"        default\u003d10,"},{"line_number":640,"context_line":"        help\u003d\"\"\""},{"line_number":641,"context_line":"Number of consecutive failed builds that result in disabing a compute service."},{"line_number":642,"context_line":""},{"line_number":643,"context_line":"This option will cause nova-compute to set itself to a disabled state"},{"line_number":644,"context_line":"if a certain number of consecutive build faulures occur. This will"}],"source_content_type":"text/x-python","patch_set":3,"id":"3f044301_6ffbbd78","line":641,"updated":"2017-05-10 03:21:48.000000000","message":"\"disabling\"","commit_id":"38fb5f3a012d7fd62a0d97ef74826b5868f49ca5"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"46cdc1a05db26d62f96fc9ebd04ff6fe7230a1f8","unresolved":false,"context_lines":[{"line_number":650,"context_line":"Number of consecutive failed builds that result in disabling a compute service."},{"line_number":651,"context_line":""},{"line_number":652,"context_line":"This option will cause nova-compute to set itself to a disabled state"},{"line_number":653,"context_line":"if a certain number of consecutive build faulures occur. This will"},{"line_number":654,"context_line":"prevent the scheduler from continuing to send builds to a compute node that is"},{"line_number":655,"context_line":"consistently failing."},{"line_number":656,"context_line":""}],"source_content_type":"text/x-python","patch_set":5,"id":"1f013ff3_0ff9557f","line":653,"range":{"start_line":653,"start_character":41,"end_line":653,"end_character":49},"updated":"2017-05-12 17:58:40.000000000","message":"spelling\n\nPlus, it would be good to note this is *all* failures, not just unknown random exceptions which trigger a retry, but also things like claims failures due to resource claim collisions. This goes back to what Chet was asking about in PS4.","commit_id":"73c180d6ee8b4ac8517ed44e1d74c1613a90bb02"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"126aac24211ff5a452292f4e0e0f50d0ce319ad0","unresolved":false,"context_lines":[{"line_number":650,"context_line":"Number of consecutive failed builds that result in disabling a compute service."},{"line_number":651,"context_line":""},{"line_number":652,"context_line":"This option will cause nova-compute to set itself to a disabled state"},{"line_number":653,"context_line":"if a certain number of consecutive build faulures occur. This will"},{"line_number":654,"context_line":"prevent the scheduler from continuing to send builds to a compute node that is"},{"line_number":655,"context_line":"consistently failing."},{"line_number":656,"context_line":""}],"source_content_type":"text/x-python","patch_set":5,"id":"1f013ff3_6e631bde","line":653,"range":{"start_line":653,"start_character":41,"end_line":653,"end_character":49},"in_reply_to":"1f013ff3_0ff9557f","updated":"2017-05-15 16:17:10.000000000","message":"Ack.","commit_id":"73c180d6ee8b4ac8517ed44e1d74c1613a90bb02"},{"author":{"_account_id":14070,"name":"Eric Fried","email":"openstack@fried.cc","username":"efried"},"change_message_id":"1a4de212390734079077cf882afd6657eec1b3ed","unresolved":false,"context_lines":[{"line_number":656,"context_line":"score, including reschedules that may have been due to racy scheduler behavior."},{"line_number":657,"context_line":"Since the failures must be consecutive, it is unlikely that occasional expected"},{"line_number":658,"context_line":"reschedules will actually disable a compute node."},{"line_number":659,"context_line":""},{"line_number":660,"context_line":"Possible values:"},{"line_number":661,"context_line":""},{"line_number":662,"context_line":"* Any positive integer representing a build failure count."}],"source_content_type":"text/x-python","patch_set":6,"id":"ff0f0b1f_25b50ace","line":659,"updated":"2017-05-19 15:53:45.000000000","message":"Perhaps this is obvious to others, but should we mention how the service can be re-enabled when the issue is resolved?","commit_id":"f93f675a8563c7c37fdcb0c685a8b491a7311361"},{"author":{"_account_id":8768,"name":"Chris Friesen","email":"chris.friesen@windriver.com","username":"cbf123"},"change_message_id":"e4c1c6902ef2ac2e8e324b76c9037163946f58c4","unresolved":false,"context_lines":[{"line_number":656,"context_line":"score, including reschedules that may have been due to racy scheduler behavior."},{"line_number":657,"context_line":"Since the failures must be consecutive, it is unlikely that occasional expected"},{"line_number":658,"context_line":"reschedules will actually disable a compute node."},{"line_number":659,"context_line":""},{"line_number":660,"context_line":"Possible values:"},{"line_number":661,"context_line":""},{"line_number":662,"context_line":"* Any positive integer representing a build failure count."}],"source_content_type":"text/x-python","patch_set":6,"id":"ff0f0b1f_173c7dbd","line":659,"in_reply_to":"ff0f0b1f_25b50ace","updated":"2017-05-19 16:11:28.000000000","message":"You use the usual \"nova service-enable\" CLI command (or the equivalent python/HTTP API).  I\u0027m not against mentioning it explicitly though.","commit_id":"f93f675a8563c7c37fdcb0c685a8b491a7311361"}],"releasenotes/notes/compute-node-auto-disable-303eb9b0fdb4f3f1.yaml":[{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"e504b6c8dcc523ec54157e0fed7c060ea16a4880","unresolved":false,"context_lines":[{"line_number":1,"context_line":"---"},{"line_number":2,"context_line":"features:"},{"line_number":3,"context_line":"  - |"},{"line_number":4,"context_line":"    Nova compute can automatically disable itself in the service database if"},{"line_number":5,"context_line":"    consecutive build failures exceed a set threshold. The"},{"line_number":6,"context_line":"    `consecutive_build_service_disable_threshold` configuration option allows"},{"line_number":7,"context_line":"    setting the threshold for this behavior, or disabling it entirely if"}],"source_content_type":"text/x-yaml","patch_set":4,"id":"3f044301_1df22471","line":4,"range":{"start_line":4,"start_character":4,"end_line":4,"end_character":17},"updated":"2017-05-10 15:03:12.000000000","message":"`nova-compute` ?","commit_id":"169576f45124d9300c81b0a6087d5c76ba30a342"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"e504b6c8dcc523ec54157e0fed7c060ea16a4880","unresolved":false,"context_lines":[{"line_number":3,"context_line":"  - |"},{"line_number":4,"context_line":"    Nova compute can automatically disable itself in the service database if"},{"line_number":5,"context_line":"    consecutive build failures exceed a set threshold. The"},{"line_number":6,"context_line":"    `consecutive_build_service_disable_threshold` configuration option allows"},{"line_number":7,"context_line":"    setting the threshold for this behavior, or disabling it entirely if"},{"line_number":8,"context_line":"    desired. The intent is that an admin will examine the issue before manually"},{"line_number":9,"context_line":"    re-enabling the service, which will avoid that compute node becoming a"}],"source_content_type":"text/x-yaml","patch_set":4,"id":"3f044301_bdec508f","line":6,"range":{"start_line":6,"start_character":4,"end_line":6,"end_character":49},"updated":"2017-05-10 15:03:12.000000000","message":"Double backticks. Also, maybe add this to a group","commit_id":"169576f45124d9300c81b0a6087d5c76ba30a342"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"e504b6c8dcc523ec54157e0fed7c060ea16a4880","unresolved":false,"context_lines":[{"line_number":2,"context_line":"features:"},{"line_number":3,"context_line":"  - |"},{"line_number":4,"context_line":"    Nova compute can automatically disable itself in the service database if"},{"line_number":5,"context_line":"    consecutive build failures exceed a set threshold. The"},{"line_number":6,"context_line":"    `consecutive_build_service_disable_threshold` configuration option allows"},{"line_number":7,"context_line":"    setting the threshold for this behavior, or disabling it entirely if"},{"line_number":8,"context_line":"    desired. The intent is that an admin will examine the issue before manually"},{"line_number":9,"context_line":"    re-enabling the service, which will avoid that compute node becoming a"},{"line_number":10,"context_line":"    black hole build magnet."},{"line_number":11,"context_line":"upgrade:"}],"source_content_type":"text/x-yaml","patch_set":4,"id":"3f044301_3d4a80ab","line":8,"range":{"start_line":5,"start_character":54,"end_line":8,"end_character":12},"updated":"2017-05-10 15:03:12.000000000","message":"Could you make this a separate paragraph?","commit_id":"169576f45124d9300c81b0a6087d5c76ba30a342"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"e504b6c8dcc523ec54157e0fed7c060ea16a4880","unresolved":false,"context_lines":[{"line_number":5,"context_line":"    consecutive build failures exceed a set threshold. The"},{"line_number":6,"context_line":"    `consecutive_build_service_disable_threshold` configuration option allows"},{"line_number":7,"context_line":"    setting the threshold for this behavior, or disabling it entirely if"},{"line_number":8,"context_line":"    desired. The intent is that an admin will examine the issue before manually"},{"line_number":9,"context_line":"    re-enabling the service, which will avoid that compute node becoming a"},{"line_number":10,"context_line":"    black hole build magnet."},{"line_number":11,"context_line":"upgrade:"},{"line_number":12,"context_line":"  - |"},{"line_number":13,"context_line":"    The new configuration option `consecutive_build_service_disable_threshold`"}],"source_content_type":"text/x-yaml","patch_set":4,"id":"3f044301_7d548853","line":10,"range":{"start_line":8,"start_character":13,"end_line":10,"end_character":28},"updated":"2017-05-10 15:03:12.000000000","message":"...and make this the second sentence","commit_id":"169576f45124d9300c81b0a6087d5c76ba30a342"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"e504b6c8dcc523ec54157e0fed7c060ea16a4880","unresolved":false,"context_lines":[{"line_number":10,"context_line":"    black hole build magnet."},{"line_number":11,"context_line":"upgrade:"},{"line_number":12,"context_line":"  - |"},{"line_number":13,"context_line":"    The new configuration option `consecutive_build_service_disable_threshold`"},{"line_number":14,"context_line":"    defaults to a nonzero value, which means multiple failed builds will"},{"line_number":15,"context_line":"    result in a compute node auto-disabling itself."}],"source_content_type":"text/x-yaml","patch_set":4,"id":"3f044301_7d02483f","line":13,"range":{"start_line":13,"start_character":34,"end_line":13,"end_character":77},"updated":"2017-05-10 15:03:12.000000000","message":"``code``","commit_id":"169576f45124d9300c81b0a6087d5c76ba30a342"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"46cdc1a05db26d62f96fc9ebd04ff6fe7230a1f8","unresolved":false,"context_lines":[{"line_number":3,"context_line":"  - |"},{"line_number":4,"context_line":"    The `nova-compute` worker can automatically disable itself in the"},{"line_number":5,"context_line":"    service database if consecutive build failures exceed a set threshold. The"},{"line_number":6,"context_line":"    ``consecutive_build_service_disable_threshold`` configuration option allows"},{"line_number":7,"context_line":"    setting the threshold for this behavior, or disabling it entirely if"},{"line_number":8,"context_line":"    desired."},{"line_number":9,"context_line":"    The intent is that an admin will examine the issue before manually"}],"source_content_type":"text/x-yaml","patch_set":5,"id":"1f013ff3_0f3ef528","line":6,"range":{"start_line":6,"start_character":6,"end_line":6,"end_character":49},"updated":"2017-05-12 17:58:40.000000000","message":"nit: [compute]consecutive_build_service_disable_threshold\n\nI think that\u0027s what sfinucan meant by \"maybe add this to a group\" in PS4.","commit_id":"73c180d6ee8b4ac8517ed44e1d74c1613a90bb02"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"46cdc1a05db26d62f96fc9ebd04ff6fe7230a1f8","unresolved":false,"context_lines":[{"line_number":11,"context_line":"    black hole build magnet."},{"line_number":12,"context_line":"upgrade:"},{"line_number":13,"context_line":"  - |"},{"line_number":14,"context_line":"    The new configuration option `consecutive_build_service_disable_threshold`"},{"line_number":15,"context_line":"    defaults to a nonzero value, which means multiple failed builds will"},{"line_number":16,"context_line":"    result in a compute node auto-disabling itself."}],"source_content_type":"text/x-yaml","patch_set":5,"id":"1f013ff3_0f271515","line":14,"range":{"start_line":14,"start_character":33,"end_line":14,"end_character":78},"updated":"2017-05-12 17:58:40.000000000","message":"``[compute]consecutive_build_service_disable_threshold``\n\nbefore sfinucan gets you for it.","commit_id":"73c180d6ee8b4ac8517ed44e1d74c1613a90bb02"}]}