Search This Blog

Friday, 29 December 2017

Verifying PCF 2.0 with PAS small footprint with bosh CLI

After installing PCF 2.0 here is how you can verify your installation using the new bosh2 CLI. In this example I use "bosh2" BUT with PCF 2.0 you can actually use "bosh". bosh2 v2 existed for a while in PCF 1.12 and some previous versions while we left bosh v1

1. SSH into your ops manager VM as shown below, in this example we using GCP

https://docs.pivotal.io/pivotalcf/2-0/customizing/trouble-advanced.html#ssh

2. Create an alias for your ENV as shown below

Note: You will need the bosh director IP address which you can obtain using

  https://docs.pivotal.io/pivotalcf/2-0/customizing/trouble-advanced.html#gather

ubuntu@opsman-pcf:~$ bosh2 alias-env gcp -e y.y.y.y --ca-cert /var/tempest/workspaces/default/root_ca_certificate
Using environment 'y.y.y.y' as anonymous user

Name      p-bosh
UUID      3c886290-144f-4ec7-86dd-b7586b98dc3b
Version   264.4.0 (00000000)
CPI       google_cpi
Features  compiled_package_cache: disabled
          config_server: enabled
          dns: disabled
          snapshots: disabled
User      (not logged in)

Succeeded

3. Log in to the BOSH Director with UAA

Note: You will need the username / password for the bosh director which you can obtain as follows

  https://docs.pivotal.io/pivotalcf/2-0/customizing/trouble-advanced.html#gather

ubuntu@opsman-pcf:~$ bosh2 -e gcp log-in
Email (): director
Password ():

Successfully authenticated with UAA

Succeeded

4. View all the VM's managed by BOSH as follows

ubuntu@opsman-pcf:~/scripts$ bosh2 -e gcp vms --column=Instance --column="Process State" --column=AZ --column="VM Type"
Using environment 'y.y.y.y' as user 'director' (bosh.*.read, openid, bosh.*.admin, bosh.read, bosh.admin)

Task 65. Done

Deployment 'cf-adee3657c74c7b9a8e35'

Instance                                             Process State  AZ                      VM Type
backup-prepare/996340c7-4114-472e-b660-a5e353493fa4  running        australia-southeast1-a  micro
blobstore/cdd6fc8d-25c9-4cfb-9908-89eb0164fb80       running        australia-southeast1-a  medium.mem
compute/2dfcc046-c16a-4a36-9170-ef70d1881818         running        australia-southeast1-a  xlarge.disk
control/2f3d0bc6-9a2d-4c08-9ccc-a88bad6382a3         running        australia-southeast1-a  xlarge
database/da60f0e7-b8e3-4f8d-945d-306b267ac161        running        australia-southeast1-a  large.disk
mysql_monitor/a88331c4-1659-4fe4-b8e9-89ce4bf092fd   running        australia-southeast1-a  micro
router/276e308e-a476-4c8d-9555-21623dada492          running        australia-southeast1-a  micro

7 vms

Succeeded

** Few other examples **

- View all the deployments, in this example we just have the PAS small footprint tile installed so it only exists and no other bosh managed tiles xist

ubuntu@opsman-pcf:~/scripts$ bosh2 -e gcp deployments --column=name
Using environment 'y.y.y.y' as user 'director' (bosh.*.read, openid, bosh.*.admin, bosh.read, bosh.admin)

Name
cf-adee3657c74c7b9a8e35

1 deployments

Succeeded

- Run cloud check to check for issues

ubuntu@opsman-pcf:~/scripts$ bosh2 -e gcp -d cf-adee3657c74c7b9a8e35 cloud-check
Using environment 'y.y.y.y' as user 'director' (bosh.*.read, openid, bosh.*.admin, bosh.read, bosh.admin)

Using deployment 'cf-adee3657c74c7b9a8e35'

Task 66

Task 66 | 04:20:52 | Scanning 7 VMs: Checking VM states (00:00:06)
Task 66 | 04:20:58 | Scanning 7 VMs: 7 OK, 0 unresponsive, 0 missing, 0 unbound (00:00:00)
Task 66 | 04:20:58 | Scanning 3 persistent disks: Looking for inactive disks (00:00:01)
Task 66 | 04:20:59 | Scanning 3 persistent disks: 3 OK, 0 missing, 0 inactive, 0 mount-info mismatch (00:00:00)

Task 66 Started  Fri Dec 29 04:20:52 UTC 2017
Task 66 Finished Fri Dec 29 04:20:59 UTC 2017
Task 66 Duration 00:00:07
Task 66 done

#  Type  Description

0 problems

Succeeded

More Information

https://docs.pivotal.io/pivotalcf/2-0/customizing/trouble-advanced.html

Tuesday, 19 December 2017

Terminating a specific application instance using it's index number in Pivotal Cloud Foundry

I was recently asked how to terminate a specific application instance rather then terminate all instances using "cf delete".

We can easily using the CF REST API or even easier the CF CLI "cf curl" command which makes it straight forward to make REST based calls into cloud foundry as shown below.

CF REST API Docs

https://apidocs.cloudfoundry.org/280/

Below assumes you already logged into PCF using the CF CLI

1. First find an application that has multiple instances

pasapicella@pas-macbook:~$ cf app pas-cf-manifest
Showing health and status for app pas-cf-manifest in org apples-pivotal-org / space development as papicella@pivotal.io...

name:              pas-cf-manifest
requested state:   started
instances:         2/2
usage:             756M x 2 instances
routes:            pas-cf-manifest.cfapps.io
last uploaded:     Sun 19 Nov 21:26:26 AEDT 2017
stack:             cflinuxfs2
buildpack:         client-certificate-mapper=1.2.0_RELEASE container-security-provider=1.8.0_RELEASE java-buildpack=v4.5-offline-https://github.com/cloudfoundry/java-buildpack.git#ffeefb9 java-main
                   java-opts jvmkill-agent=1.10.0_RELEASE open-jdk-like-jre=1.8.0_1...

     state     since                  cpu    memory           disk           details
#0   running   2017-12-16T00:11:27Z   0.0%   241.5M of 756M   139.9M of 1G
#1   running   2017-12-17T10:39:09Z   0.3%   221.3M of 756M   139.9M of 1G

2. Use a "cf curl" curl which uses the application GUID to determine which application to check all application instances and their current state

pasapicella@pas-macbook:~$ cf curl /v2/apps/`cf app pas-cf-manifest --guid`/instances
{
   "0": {
      "state": "RUNNING",
      "uptime": 293653,
      "since": 1513383087
   },
   "1": {
      "state": "RUNNING",
      "uptime": 169591,
      "since": 1513507149
   }
}

3. Now let's delete instance with index "1". Don't forget that PCF will determine the current desired state of the application is not the current state and will re-start the application instance very quickly

pasapicella@pas-macbook:~$ cf curl /v2/apps/`cf app pas-cf-manifest --guid`/instances/1 -X DELETE

Note: You won't get any output BUT you can verify it has done what you asked for by running the command at step #2 again

pasapicella@pas-macbook:~$ cf curl /v2/apps/`cf app pas-cf-manifest --guid`/instances
{
   "0": {
      "state": "RUNNING",
      "uptime": 293852,
      "since": 1513383087
   },
   "1": {
      "state": "DOWN",
      "uptime": 0
   }
}

If you run it again say 30 seconds later you should see your application instance re-started as shown below

pasapicella@pas-macbook:~$ cf curl /v2/apps/`cf app pas-cf-manifest --guid`/instances
{
   "0": {
      "state": "RUNNING",
      "uptime": 293870,
      "since": 1513383087
   },
   "1": {
      "state": "STARTING",
      "uptime": 11,
      "since": 1513676947
   }
}

pasapicella@pas-macbook:~$ cf curl /v2/apps/`cf app pas-cf-manifest --guid`/instances
{
   "0": {
      "state": "RUNNING",
      "uptime": 293924,
      "since": 1513383087
   },
   "1": {
      "state": "RUNNING",
      "uptime": 45,
      "since": 1513676965
   }
}

More Information

pasapicella@pas-macbook:~$ cf curl --help
NAME:
   curl - Executes a request to the targeted API endpoint

USAGE:
   cf curl PATH [-iv] [-X METHOD] [-H HEADER] [-d DATA] [--output FILE]

   By default 'cf curl' will perform a GET to the specified PATH. If data
   is provided via -d, a POST will be performed instead, and the Content-Type
   will be set to application/json. You may override headers with -H and the
   request method with -X.

   For API documentation, please visit http://apidocs.cloudfoundry.org.

EXAMPLES:
   cf curl "/v2/apps" -X GET -H "Content-Type: application/x-www-form-urlencoded" -d 'q=name:myapp'
   cf curl "/v2/apps" -d @/path/to/file

OPTIONS:
   -H            Custom headers to include in the request, flag can be specified multiple times
   -X            HTTP method (GET,POST,PUT,DELETE,etc)
   -d            HTTP data to include in the request body, or '@' followed by a file name to read the data from
   -i            Include response headers in the output
   --output      Write curl body to FILE instead of stdout