Skip to content

Latest commit

 

History

History
103 lines (80 loc) · 3.77 KB

README.md

File metadata and controls

103 lines (80 loc) · 3.77 KB
Raw

Table of Content

Use-Case - Engine Event Triggering

The idea:

  • To show how we monitor a Junos string within a file and convert it to an event.
    • Natively this not supported within SaltStack; thereby, included in this use-case a custom beacon module.

Example show here:

  • Monitor ERROR on a salt-minion file called /var/log/random.log and upon receiving this ERROR from the minion on the master, master restarts the process and writes a file in /tmp/random_process_restart.log

The problem which it solves:

  • Consider a topology where devices are getting provisioned based on a dynamic process e.g. ZTP. If one is able to monitor any log with any regex, one can detect a malfunction and restart the process e.g. for ZTP the DHCP server.

The components used for the example:

  • Custom beacon module for SaltStack called log.py programmed by Iddo Cohen.
  • salt-master called saltmaster-beacon used to sync beacon configuration to minions.
  • salt-minion called minion01 having /var/log/random.log

Spin-up the use-case

All commands must be executed under main directory (cd ..)

0- Get/build image

Get the image via docker hub:

docker pull juniper/saltstack

or build your own image from git:

make build

1- Start use-case

To start the use-case:

make start-uc-beacon

Which is the same as:

make master-start UC='beacon'
make minion-start DEVICE='minion01' UC='beacon'
(wait X seconds until minion starts)
make accept-keys UC='beacon'
(wait X seconds until master registers minion)
(execute on the saltmaster-beacon this command 'salt "minion01" saltutil.sync_beacons')
docker restart minion01

2- Start minion-shell and test

To test the event handling, start the minion01 shell.

make minion-shell DEVICE='minion01' UC='beacon'

Within the shell, check content of /var/log/random.log:

root@minion01:/# cat /var/log/random.log
Started daemon

Also check if /tmp/random_process_restart.log does not exists:

root@minion01:/tmp# ls -la
total 8
drwxrwxrwt  3 root root 4096 Nov 20 12:26 .
drwxr-xr-x 74 root root 4096 Nov 20 12:22 ..

Test the event by echoing ERROR message into random.log:

root@minion01:/# echo "Critical ERROR occured" >> /var/log/random.log

And check if the new file /tmp/random_process_restart.log with the content has been created:

root@minion01:/tmp# cat random_process_restart.log
Restarting random deamon because of error: Critical ERROR occured

Under the hood

  • At start-up (make start-uc-beacon):
    • A reactor is configured on the saltmaster-beacon under /etc/salt/master.d/reactor.conf which gets executed and calls a rule set under /srv/reactor/monitor_message.sls
    • The customer beacon is put under /srv/salt/_beacon/ on the saltmaster-beacon and syncronised to minion01 via salt "minion01" saltutil.sync_beacons
    • Because of ticket # the SaltStack minion minion01 needs to be restarted. This is done via docker restart minion01
  • At operation:
    • Under the /srv/pillar/top.sls on the saltmaster-beacon a beacon is defined under /srv/pillar/log.sls. This definition is the also syncronised to minion01, so the minion knows what needs to be done.
    • Within the log.sls there is a configuration specifying a regex which should be matched.
    • When a match has been found, the raw message is send to the reactor on the event-bus to the saltmaster-beacon, which in return executes /srv/reactor/monitor_message.sls