Simplified `cron` deployment using Sparrow as Jekyll server

[siddhartha-gadgil]

A better way to manage deployment may be to use sparrow as a jekyll server. Individual steps have been tested and timed (total time less than a minute), but scripts are not yet written.

On Sparrow

• A script which (running with cron)
  • pulls from the repository
  • builds with jekyll for deployment
  • creates a tarball (i.e., zips)
  • copies tarball to an appropriate directory, say zipsite.
• run in the background an http-server to serve zipsite - the node-js http-server is installed and I have used it many times.

On math

A script to:

• Download the tarball from Sparrow
• extract the tarball
• recursively copy to the running site.

Running and cron

• Both Sparrow and math will have scripts which can be triggered manually and will also run as cron jobs.
• the script on math should be run as root.

[siddhartha-gadgil]

• scripts are written and tested a little.
• no cron jobs yet.
• jekyll server running on sparrow.

[siddhartha-gadgil]

• Everything tested. The cron job on math, for a deployment every half hour is:

  SHELL=/usr/bin/bash
  */2 * * * * /home/gadgil/ziploy.sh /tmp/webpage &>> /tmp/ziploy.log

• the /tmp/webpage should be replaced by the correct target or use default.
• the cron job on sparrow is:

  SHELL=/bin/bash
  0 * * * * /home/gadgil/code/DeptWeb/mktar.sh &>> /home/gadgil/mktar.log

[ayyer]

Ziploy is a nice name!

[siddhartha-gadgil]

Many attempts have failed. Instead there is simplified deployment from sparrow.