cfk

0.3.7 / 2016-04-20 ================== * add deploy/sync commands for chronos tasks 0.3.6 / 2016-04-16 ================== * add port envvars to sync 0.3.5 / 2016-04-16 ================== * actually fix environment variable escaping 0.3.4 / 2016-04-16 ================== * fix environment variable escaping 0.3.3 / 2016-04-16 ================== * upgrade mesos & marathon 0.3.2 / 2016-04-15 ================== * significantly improve syncing process * exit right away if no cluster is specified 0.3.1 / 2016-04-04 ================== * fix service checks 0.3.0 / 2016-04-04 ================== * prompt for consul configuration options * router: deduplicate haproxy template rendering * router: support specification of required services via environment * router: configurable max-stale * router: upgrade to consul-template v0.14.0 0.2.1 / 2016-04-01 ================== * try a different keyserver 0.2.0 / 2016-04-01 ================== * upgrade mesos/marathon/chronos 0.1.12 / 2016-03-27 =================== * add configurable min/max wait for consul-template 0.1.11 / 2016-03-06 =================== * fix cfk-home path on initial sync 0.1.10 / 2016-03-06 =================== * make names unique to support multiple clusters under one account. use default ebs_optimized setting for instance type * set id_rsa permissions on cluster sync * fix getopts on cfk run * move env.upgrade on init and fix permissions of id_rsa on sync * only set default environment variables if not 'init' 0.1.9 / 2016-02-22 ================== * add haproxy check to avoid conflicting ssl options for root domains with matching tlds 0.1.8 / 2016-02-22 ================== * fix dns health check to check elb instead of domain 0.1.7 / 2016-02-22 ================== * fix route53 dns record configuration for primary sets * cluster support for fallback url (secondary failover records must be configured manually) * add router support for fallback url * fix calling symlinked subcommands * fix calling symlinked subcommands 0.1.6 / 2016-02-19 ================== * support multiple root domains 0.1.5 / 2016-02-12 ================== * add gettext to install script (for envsubst) 0.1.4 / 2016-02-11 ================== * fix cfk init password location 0.1.3 / 2016-02-11 ================== * fix installation bugs, prefix subcommands with cfk- 0.1.2 / 2016-02-11 ================== * add brew upgrade to install 0.1.1 / 2016-02-11 ================== * add some more instruction to installation script 0.1.0 / 2016-02-11 ================== * basic install script (mac only) 0.0.114 / 2016-02-08 ==================== * disable elk stack for now 0.0.113 / 2016-02-08 ==================== * add custom domain support * combine app stdout and stderr 0.0.112 / 2016-02-08 ==================== * support tailing marathon app logs * include only variables defined in dockerfile as build-args 0.0.111 / 2016-02-08 ==================== * don't upgrade kernel (docker bug with aufs + kernel>3.13.0-71 * support adding env to docker build 0.0.110 / 2016-02-07 ==================== * fix router regex match 0.0.109 / 2016-02-07 ==================== * always deregister containers from consul on exit regardless of status 0.0.108 / 2016-02-07 ==================== * ensure unique server names in haproxy by including partial container id for mesos tasks * fix auto-added ports on deploy 0.0.107 / 2016-02-06 ==================== * add port info to marathon demo config * change default port to 80, add all ports to env as PORT{N} * add CFK_ prefix to marathon environment variables 0.0.106 / 2016-02-06 ==================== * create pid tracking file for log processes in /tmp instead of current directory * fix subcommand list * switch to aufs driver 0.0.105 / 2016-02-05 ==================== * pin docker to 1.9.1, weave to 1.4.3, scope to 0.12.0, discovery to 0.1 0.0.104 / 2016-02-05 ==================== * docker 1.10 requires libsystemd-journal, so need to link into agent-mesos 0.0.103 / 2016-02-05 ==================== * fix discovery monitor with new container name * explicitly make shyaml executable 0.0.102 / 2016-02-05 ==================== * fix WEAVE_DISCOVERY_NAME 0.0.101 / 2016-02-05 ==================== * fix problem with discovery & docker 1.10 (IMAGE_VERSION not defined) 0.0.100 / 2016-02-05 ==================== * add docker-login, add safe sync option for unattended sync, read password into env * simplify weave peer detection command * forget unreachable peers instead of rm * use local ip for consul template * fix s3 sync after upgrade 0.0.99 / 2016-02-03 =================== * fix weave-consul sync 0.0.98 / 2016-02-03 =================== * fix weave consul sync 0.0.97 / 2016-02-03 =================== * fix weave-consul sync * fix auto rollup on apply 0.0.96 / 2016-02-03 =================== * update consul precheck since join is dynamic now 0.0.95 / 2016-02-03 =================== * fix elasticsearch data path * add stability interval to upgrade script 0.0.94 / 2016-02-02 =================== * revert resource constraints in packer json 0.0.93 / 2016-02-02 =================== * add color to hosts command, fix bug in logs * check and remove dead consul peers 0.0.92 / 2016-02-02 =================== * add support for mounting ephemeral storage * move plan/apply responsibility to upgrade from config * move large volumes to /data directory on host 0.0.91 / 2016-02-01 =================== * add apply/upload to upgrade script * upgrade script for rolling configuration/version updates 0.0.90 / 2016-02-01 =================== * move config into separate command so existing clusters can be easily updated * fix consul peer join 0.0.89 / 2016-02-01 =================== * add cfk admin command line tools * add X-Forwarded-Proto header to router traffic * split dns and ssl into terraform overrides files (for templated cluster creation) * fix bug reporting free disk space warning * restrict elasticsearch to master hosts only and add periodic consul-join task to keep in sync with weave peers 0.0.88 / 2016-01-23 =================== * fix haproxy.cfg.tmpl error 0.0.87 / 2016-01-23 =================== * fix ip grep on weave connections * put master and agent on same ip range when running in vagrant * fix upgrade script wait 0.0.86 / 2016-01-23 =================== * fix weave peer list 0.0.85 / 2016-01-23 =================== * enable leader discovery without constant calls to weave status (hangs docker) * only use s3 for zookeeper if explicitly configured * add route for cfk-role * Add instance profile to agent asg 0.0.84 / 2016-01-22 =================== * remove aws access key variables (replaced by IAM policy) * upgrade consul version * add s3 bucket and access policies to terraform * use host ip instead of hostnames for consul health checks, choose consul leader for bootstrapping based on sorted master ips * better formatting for remote log output * remove router.local from default search domain...needs to be explicitly requested because it implies the assumption that service is on port 80 (not the case when using weaveDNS or consulDNS ip) 0.0.83 / 2016-01-21 =================== * add upgrade command to makefile * try fixing zookeeper ensemble size and increasing reaper timeout * updates to upgrade script, add run script for one-off ssh commands across hosts 0.0.82 / 2016-01-21 =================== * add bin scripts, tag elb, and turn off upstart script debugging 0.0.81 / 2016-01-20 =================== * add simple http endpoint for checking platform version 0.0.80 / 2016-01-19 =================== * fix pip pin 0.0.79 / 2016-01-19 =================== * pin pip version (version 8.0.0 released today has some issues) 0.0.78 / 2016-01-19 =================== * make slack reports more descriptive, easier to distinguish * add slack startup/shutdown notifications * add resource limits to packer builder * add monitor to check and remove dead peers from weave 0.0.77 / 2016-01-19 =================== * pass cfk_version into image build to fix initial docker pull 0.0.76 / 2016-01-19 =================== * integrate glue containers into main project (router, reaper, reporter) * add cfk version to environment * fix bug in scope_enabled detection * add host/container monitoring slack alerts * add private marathon docker app definition 0.0.75 / 2016-01-16 =================== * expose router port 81 for proxy protocol 0.0.74 / 2016-01-16 =================== * add support for proxy protocol from ELB (to enable https redirect while maintaining websocket support on aws) 0.0.73 / 2016-01-16 =================== * increase agent timeout * trigger rebuild to get latest reaper 0.0.72 / 2016-01-16 =================== * fix user data variables on ec2 * better preconditions for mesos, marathon, and chronos * fix incorrect resetting of reaper failed counts 0.0.71 / 2016-01-15 =================== * keep vagrant input artifact 0.0.70 / 2016-01-15 =================== * fix quotes in userdata script * expose weave network regardless of whether scope is enabled 0.0.69 / 2016-01-15 =================== * fix syntax error in packer template 0.0.68 / 2016-01-15 =================== * add ssl + basic auth options to terraform * add apt-get upgrade to image build * put core services behind basic auth * move health check url and add support for basic auth + ssl redirect * add admin user/pass and ssl config to env * move configuration options out of vagrantfile * make scope a configurable service * fix NUM_MASTERS on agent nodes 0.0.67 / 2016-01-15 =================== * fix consul check pre-conditions on agent nodes 0.0.66 / 2016-01-15 =================== * fix docker-cleanup file path 0.0.65 / 2016-01-15 =================== * fix mesos agent zookeeper url * clean up potentially corrupted local-kv.db file pre docker start 0.0.64 / 2016-01-15 =================== * fix agent-mesos precondition 0.0.63 / 2016-01-14 =================== * change packer config back. 0.0.62 / 2016-01-14 =================== * update master/agent count for test * fix mesos health check, pin consul to 0.5 (0.6 does not work\!) * fix INITIAL_PEERS count * Reorder packer artifacts following Alvaro's advice * fix weave environment * Thanks Alvaro! 0.0.61 / 2016-01-14 =================== * fix default docker opts to use host ip for default dns (cfk env file changed names) * restart docker at end of provisioning 0.0.60 / 2016-01-14 =================== * fix userdata script in terraform and QUORUM_NUM string->int conversion * add builder-specific make commands for more efficient local packing 0.0.59 / 2016-01-14 =================== * get rid of useless docker check 0.0.58 / 2016-01-13 =================== * major restructuring * much improved health checks * preconditions to prevent premature starts that cause bootstrapping to fail 0.0.57 / 2016-01-12 =================== * exit from post-start script if respawn requested 0.0.56 / 2016-01-12 =================== * abandon attempts to add local dns entries (creates too many containers) * improve startup order using post-start to delay started event * fix start-after ordering bug in agent upstart tasks * start/stop weave once (instead of restarting every service on provisioning * add single dns monitor for all services to prevent spawning tons of weave containers * remove remnants of old attempt to prevent simultaneous container starts * add more cpus to vagrant boxes 0.0.55 / 2016-01-11 =================== * set default docker option to use only router dns * add dns settings to discovery * add dns check to service monitors * fix dnsmasq.hosts newlines 0.0.54 / 2016-01-11 =================== * fix chronos/marathon health checks * update marathon/chronos framework names 0.0.53 / 2016-01-10 =================== * add health check to ensure chronos and marathon are registered with mesos 0.0.52 / 2016-01-10 =================== * try different health-check grace periods * add make command to teardown asg * remove debug from reaper 0.0.51 / 2016-01-10 =================== * add reaper container to restart failing containers * fix tf.apply destruction of changed asgs 0.0.50 / 2016-01-09 =================== * fix makefile * better chronos healthcheck 0.0.49 / 2016-01-09 =================== * fix makefile 0.0.48 / 2016-01-09 =================== * have tf.apply destroy previous autoscaling groups if launch configurations have changed (need to do this on a rolling basis eventually) * cleanup spacing in userdata * restore monitor failure rule * move packer cache dir and add terraform shortcuts to makefile 0.0.47 / 2016-01-09 =================== * scale example marathon app to 2 instances by default * shorten vhost urls for mesos services 0.0.46 / 2016-01-09 =================== * packer variable: roles => role 0.0.45 / 2016-01-09 =================== * Adds MIN_AGENTS and CFK_ROLE to cfk environment * Fixes weave -init-peer-count to expect a minimum number of agents * Tell registrator to keep services registered in consul when they die unexpectedly * Improve templating of haproxy health monitoring rules * Add consul healthchecks for zookeeper, consul, chronos, marathon, elasticsearch, and kibana * Fixes marathon/mesos/chronos web_ui urls * Fixes hostname of registrator to match server hostname * Fixes mesos-slave containerizers * Adds scope to the haproxy health url * Adds a sample marathon app definition * Reset dnsmasq cache when hosts file changes 0.0.44 / 2016-01-06 =================== * improve mesos slave health check 0.0.43 / 2016-01-06 =================== * escape $ in docker-compose.yml 0.0.42 / 2016-01-06 =================== * add retry to scope host verification healthcheck * fix mesos slave healthcheck to ensure it's registered with master * add scope healthcheck to restart if host not registered * explicitly use weave dns for mesos zookeeper dns 0.0.41 / 2016-01-06 =================== * use zookeeper for mesos agent leader detection 0.0.40 / 2016-01-05 =================== * fix haproxy backends * distinguish between agent and master http services for better haproxy templating 0.0.39 / 2016-01-05 =================== * shorten health check grace period * remove consul agent from haproxy, add healthcheck to ensure unregistered slaves are terminated 0.0.38 / 2016-01-05 =================== * force registrator to use local consul container 0.0.37 / 2016-01-05 =================== * add mesos healthchecks to agent haproxy instances * fix advertised hostnames for zk, mesos, marathon, chronos 0.0.36 / 2016-01-04 =================== * add agent infrastructure * add server metadata to aggregated logs 0.0.35 / 2016-01-03 =================== * finish renaming cfk-agent env file * remove scope token * change host-specific weave dns entries to - instead of . subdomain * rename cfk-agent env file * force remove dangling weave dns entries on start/stop * rename mesos-agent to agent-mesos (mesos-* prefix is reserved) 0.0.34 / 2016-01-02 =================== * add jq and fix mesos consul health check * add agent image 0.0.33 / 2016-01-01 =================== * add consul health check to mesos via registrator env vars 0.0.32 / 2016-01-01 =================== * switch to ansible_local for vagrant provisioning * fix elasticsearch discovery 0.0.31 / 2015-12-31 =================== * add haproxy health checks to elb * fix chronos config 0.0.30 / 2015-12-31 =================== * fix chronos config and haproxy master health check 0.0.29 / 2015-12-31 =================== * upgrade to latest, official mesos (0.26.0) marathon (0.13.0) chronos (2.4.0) images 0.0.28 / 2015-12-31 =================== * fix haproxy health check to fail mesos non-leader backends 0.0.27 / 2015-12-31 =================== * fix haproxy command and add default unhealthy acl 0.0.26 / 2015-12-30 =================== * fix scope/discovery monitoring jobs 0.0.25 / 2015-12-30 =================== * fix monitor script 0.0.24 / 2015-12-30 =================== * add scope & discovery monitor config to provisioning, fix error in service monitor template 0.0.23 / 2015-12-30 =================== * add lost container check, monitor scope and discovery 0.0.22 / 2015-12-30 =================== * fix stall counter 0.0.21 / 2015-12-30 =================== * fix greaser script template * revert to restarting all services instead of just weave * enable debug mode for post-start and pre-stop service scripts * add greasers to monitor stuck upstart jobs * don't try to rm all containers on weave startup * add debugging for hanging upstart jobs * only restart weave service on instance provisioning 0.0.20 / 2015-12-29 =================== * copy initial haproxy config on ansible build 0.0.19 / 2015-12-29 =================== * add default haproxy config and fix consul url envvar in router 0.0.18 / 2015-12-29 =================== * fix local packer build command, fix size of launch ami * increase ami builder volume size 0.0.17 / 2015-12-29 =================== * add ELK stack for log aggregation, more stable service initialization 0.0.16 / 2015-12-15 =================== * fix scope & service dns entry attempts 0.0.15 / 2015-12-15 =================== * fix QUORUM_NUM calculation 0.0.14 / 2015-12-15 =================== * fix defaults exports, update vagrant provisioning in vagrantfile, aws provisioning with userdata 0.0.13 / 2015-12-15 =================== * fix packer paths 0.0.12 / 2015-12-15 =================== * fix packer ansible paths 0.0.11 / 2015-12-15 =================== * split provisioners and add initial terraform config 0.0.10 / 2015-12-12 =================== * test vagrant update 0.0.9 / 2015-12-12 ================== * more sensible atlas uploads * remove atlas post-processors on dirty builds 0.0.8 / 2015-12-12 ================== * setup ssh for vagrant box 0.0.7 / 2015-12-12 ================== * only apply post-processors (atlas upload) to clean builds 0.0.6 / 2015-12-12 ================== * fix vbox ssh/sudo password 0.0.5 / 2015-12-12 ================== * remove password variable 0.0.4 / 2015-12-12 ================== * remove root password for vagrant, add ubuntu user 0.0.3 / 2015-12-12 ================== * fix docker pull * add push to packer, restructure files 0.0.2 / 2015-12-10 ================== * update makefile dirty checks 0.0.1 / 2015-12-10 ================== * Initial Commit