Skip to content

Latest commit

 

History

History
216 lines (182 loc) · 9.12 KB

README_supplemental.md

File metadata and controls

216 lines (182 loc) · 9.12 KB

This page provides supplementary information for Automate-UI. See the main README for the larger picture.

Sections

Config Management Service Notes

Always refer to the Config Management Service repo for the most up-to-date information that may supercede anything here at any time :-)

Introduction

Once you pull down the config-mgmt service repository and setup its prerequisites as described in the README.md for the service, navigate to its directory. If you have direnv wired up correctly you will see several environment variables beginning with HAB being exported:

$ cd config-mgmt-service
direnv: loading .envrc
direnv: export +HAB_DOCKER_OPTS +HAB_ORIGIN +HAB_UPDATE_STRATEGY_FREQUENCY_MS

The most common reason that the env variables fail to appear is that the direnv setup (eval "$(direnv hook bash)") really needs to be the last item in your .bashrc or .bash_profile.

The sequence of steps to stand up the config-mgmt service--as revealed by the dev_help command in habitat studio--is just:

# setup_dev_environment
# build
# start

Note that you will still have no data in the service yet; how to do that is described near the end of this page.

The Build Step

A successful build should conclude with something like this:

. . .
» Signing config-mgmt-service_blake2bsums
☛ Signing config-mgmt-service_blake2bsums with MyUserName-20170905215937 to create /hab/pkgs/MyUserName/config-mgmt-service/0.1.0/20170906171418/FILES
★ Signed artifact /hab/pkgs/MyUserName/config-mgmt-service/0.1.0/20170906171418/FILES.
   config-mgmt-service: Generating package artifact
. . .
★ Signed artifact /hab/cache/artifacts/MyUserName-config-mgmt-service-0.1.0-20170906171418-x86_64-linux.hart.
'/hab/cache/artifacts/MyUserName-config-mgmt-service-0.1.0-20170906171418-x86_64-linux.hart' -> '/src/results/MyUserName-config-mgmt-service-0.1.0-20170906171418-x86_64-linux.hart'
   config-mgmt-service: hab-plan-build cleanup
   config-mgmt-service:
   config-mgmt-service: Source Path: /src
   config-mgmt-service: Installed Path: /hab/pkgs/MyUserName/config-mgmt-service/0.1.0/20170906172232
   config-mgmt-service: Artifact: /src/results/MyUserName-config-mgmt-service-0.1.0-20170906172232-x86_64-linux.hart
   config-mgmt-service: Build Report: /src/results/last_build.env
   config-mgmt-service: SHA256 Checksum: d3136a649d9c91cd78c55fd879fbdf0d0ae006dc11294cf0cf4fc80f394efb98
   config-mgmt-service: Blake2b Checksum: 5a425312ec68413ab9af1ec51c92992f2e3b117b4a59495178a37a9850f62be3
   config-mgmt-service:
   config-mgmt-service: I love it when a plan.sh comes together.
   config-mgmt-service:
   config-mgmt-service: Build time: 0m57s
[default:/src:0]#

Besides the above log output you can confirm that the build was successful by invoking the actual service binary directly. From the log trail above, find the line beginning with Signed artifact and ending with FILES. Take that path and replace the leaf FILES with bin. Navigate to that directory, where there should be a config-mgmt-service executable and just run it. This should result in a brief man page for the service. So /hab/pkgs/MyUserName/config-mgmt-service/0.1.0/20170906171418/FILES becomes /hab/pkgs/MyUserName/config-mgmt-service/0.1.0/20170906171418/bin and then:

[default:/src:0]# cd /hab/pkgs/MyUserName/config-mgmt-service/0.1.0/20170906171418/bin
[default:/src:0]# ./config-mgmt-service

NAME:
   config-mgmt-service - config-mgmt API Service

USAGE:
   config-mgmt-service [global options] command [command options] [arguments...]

VERSION:
   0.1.0

COMMANDS:
     start    Starts the config-mgmt API service
     help, h  Shows a list of commands or help for one command

GLOBAL OPTIONS:
   --help, -h     show help
   --version, -v  print the version

COPYRIGHT:
   Copyright 2017, Chef Software Inc.

Starting the config-mgmt service

What you should see when you start the config-mgmt service is something like this:

[default:/src:0]# start
hab-sup(MN): Supervisor starting MyUserName/config-mgmt-service. See the Supervisor output for more details.
 => Waiting to load svc MyUserName/config-mgmt-service
 => Waiting to load svc MyUserName/config-mgmt-service
. . .
» Applying configuration for config-mgmt-service.default incarnation 1504718624
Ω Creating service configuration
✓ Verified this configuration is valid TOML
↑ Applying to peer 127.0.0.1:9638
★ Applied configuration
 => Waiting for port 1234 to be listening
 => Waiting for port 1234 to be listening
[default:/src:0]#

To troubleshoot starting the service, it may be helpful to monitor the supervisor log (sup-log or sl for short) while the service attempts to start up. It is cleaner to do this in a separate window, of course, but that means you would be outside habitat studio. sup-log just runs tail on the supervisor's log file and this can easily be done externally with just one extra step. First, identify the name of your docker instance (right-most column below). Note that this name is a random string, and is regenerated each time you start the config-mgmt service.

$ docker ps
CONTAINER ID  IMAGE                     COMMAND                 CREATED         STATUS      PORTS                      NAMES
b1fd6b4d036d  habitat.../studio:0.30.2  "/bin/hab studio e..."  10 minutes ago  Up 10 mins  0.0.0.0:1234->1234/tcp...  determined_ritchie

Supply the above name to this next command to see the supervisor log outside of habitat studio:

$ docker exec -it determined_ritchie tail -f /hab/studios/src/hab/sup/default/sup.log

→ Using core/libsodium/1.0.11/20170514010744
→ Using core/linux-headers/4.3/20170513200956
→ Using core/openssl/1.0.2j/20170513215106
→ Using core/xz/5.2.2/20170513214327
→ Using core/zlib/1.2.8/20170513201911
✓ Installed core/hab-launcher/4571/20170727193541
★ Install of core/hab-launcher/4571/20170727193541 complete with 1 new packages installed.
hab-sup(MR): Supervisor Member-ID 6b15ce72b3664e8f9b8d13174b83e704
hab-sup(MR): Starting gossip-listener on 0.0.0.0:9638
hab-sup(MR): Starting http-gateway on 0.0.0.0:9631
hab-sup(PK): Confirmed latest version of core/elasticsearch is core/elasticsearch/5.5.0/20170726170852
hab-sup(MR): Starting core/elasticsearch
hab-sup(PK): Confirmed latest version of core/elasticsearch is core/elasticsearch/5.5.0/20170726170852
elasticsearch.default(SR): Hooks recompiled
. . .
hab-sup(MR): Starting MyUserName/config-mgmt-service
config-mgmt-service.default(SR): Hooks recompiled
. . .

Should you want to restart the config-mgmt service just stop and restart it like so:

[default:/src:0]# hab svc stop $HAB_ORIGIN/$pkg_name
[default:/src:0]# start

(Rumor has it that there will soon just be a simple stop command available.)

If the service is not starting up properly, also check the running processes inside habitat studio. You should see something similar to this:

[default:/src:0]# ps

PID   USER  TIME COMMAND
    1 root  0:00 {hab-studio} /bin/sh /hab/pkgs/core/hab-studio/0.30.2/20170822233239/bin/hab-studio enter
  132 root  0:00 /hab/pkgs/core/bash/4.3.42/20170513213519/bin/bash --login +h
 3766 root  0:00 /hab/pkgs/core/bash/4.3.42/20170513213519/bin/bash --login +h
 3784 root  0:01 /hab/pkgs/core/hab-launcher/4571/20170727193541/bin/hab-launch start core/elasticsearch
 3787 root  0:11 /hab/pkgs/core/hab-sup/0.30.2/20170822230509/bin/hab-sup start core/elasticsearch
 3855 hab   0:52 /hab/pkgs/core/jre8/8u131/20170622181030/bin/java -Xms1g -Xmx1g -XX:+UseConcMarkSweepGC
 5144 hab   0:00 config-mgmt-service start --port 1234 --host 0.0.0.0 --backend elasticsearch --elasticsearch
 5294 root  0:00 ps

If not, you could try rerunning the setup command (setup_dev_environment) then invoking start. (You do not need to rerun the build step if you have not left habitat studio.)

Once the config-mgmt service has successfully started you should be able to ping it:

$ curl http://localhost:1234/version
{"SHA":"9703cd1","built":"2017-09-06_17:22:49","name":"config-mgmt-service","version":"0.1.0"}

Generating Sample Nodes

There will, however, be no nodes available with that minimum setup. Add this next command to generate 100 nodes with a random distribution of statuses:

[default:/src:0]# generate_elasticsearch_data
. . .
wrote node name Qui-quis-illum with id 97b5f6bb-cf7d-4f18-966d-1f70c94acd7e
wrote node name Consectetur-nisi-possimus with id 76b2c016-e19e-4bd1-83d5-e0b636c4c864
wrote node name facere-saepe-esse with id 9d3c5159-7d57-486a-a248-8823474b9edc
wrote node name Est-possimus-qui with id 49871fc7-b520-415e-a87b-d107640ba934
wrote node name totam-nobis-at with id 6bdddcce-0086-4b43-b2dc-a6379b0cdb3f
wrote node name accusamus-quo-et with id 0f56463b-48fc-43c9-912f-55cba36d6103
. . .

(Note that this command should be run after the build but it could be before or after you start the service.)

You can then query further:

$ curl http://localhost:1234/stats/node_counts
{"success":35,"failure":32,"missing":33,"total":100}