Difference between revisions of "Project Bringing Sexy Back"

From the Linux and Unix Users Group at Virginia Teck Wiki
Jump to: navigation, search
imported>Echarlie
(Getting new hardware)
imported>Echarlie
Line 5: Line 5:
 
Oh, and tshirts.
 
Oh, and tshirts.
  
//TODO talk about doing stuff from the project list
+
<!--TODO talk about doing stuff from the project list-->
  
  
[[T-shirts]] //TODO update with designs
+
[[T-shirts]] <!--TODO update with designs-->
  
  

Revision as of 07:37, 1 December 2017

This is an effort toward the end of Fall 2017 to get new hardware, determining if we can use old hardware for anything, bring many vtluug services back to life plus add new services, fix major problems relating to the current infrastructure, make managing the current infrastructure easier, and to provide enough documentation so that anybody can have a very good understanding of how everything works without having to ask the officers about every little everything. This goes beyond documentating what we have (aka the wiki), instead provide a sort of guide that provides an overview on how to run everything. Yes, we realize this is a massive task. Timeline TBD.

One major change we are considering is running an OpenStack cluster for managing everything, along with offering a VPS service in addition to shell accounts on Acidburn. This is not possible with the current hardware.

Oh, and tshirts.


T-shirts


Getting new hardware

Cyberdelia is the only connected physical machine right now, besides our router. See Infrastructure 2017. We are planning to request funding from the SBB for new infrastructure as early as possible Spring 2018.

We are still working on exactly what to ask for, but here are some desirable things: //TODO link items

Priority Item Estimated Cost
10 Dell PowerEdge R430 $1,200 - $1,400
10 RAM (R430s only have 8GB) ~ $245/16GB
9 1Gbps MANAGED Switch ~ $100
8 UPS TBD - Based on infra power usage.

Old hardware evaulation

Wood

Phantomphreak and Joey

Infrastructure Organization

Planned Infrastructure Organization

  • Use Debian instead of Centos due to better compataiblity with ZFS and upgradability. Regarding the effectively dead EL Migration Project, I'd like to see some specific examples of this if it's an issue because both Debian and Centos are both very popular for servers.
  • Continue using temp88191 as a router. OS TBD. Currently running ubuntu 14.04 (upgrade to 16.04 possible) Also see our networking setup
  • Get a new compute server, and use Cyberdelia as a NAS (possibly using FreeNAS or FreeBSD).
  • Possibly use some old hardware for part of an OpenStack cluster.

Planned Services

REQUIRED

  • foodforus
  • Improved homepage using bootstrap and some webshit.

VERY IMPORTANT

IMPORTANT

Other things to consider

  • diaspora*
  • finger (kerberos)
  • talkd
  • member web pages in vtluug.org/~username/

Issues with current services

For all services using a custom systemd script

  • Need standard place to put script for organization (i.e. some git repo)
    • Wadsworth contains debian SysV init scripts in the ircbot homedir
    • VTBash contains scripts in pew's homedir (which echarlie moved to a sane place)
    • router configs are on github, as are old ldap configs and triggered in rc.local

VTLUUG wiki

  • Need meeting script
  • Some meetings weren't added to category (fixed??)
  • Lots of out of date information
  • Templates?

Gobblerpedia

  • Lots of out of date information
  • Need templates for buildings---currently using bloated wikipedia scripts that don't reflect our needs
  • automagically add data scraped from webpages to gobblerpedia

VTBash

  • bash.vtluug.org is working as of August 2017
  • An issue with the code resulted in 2 different databases, eash of which is very large and good
    • We need to convert these sqlite dbs
  • Needs overhaul
    • Shittly converted to python 3 from python 2 and wordwrap isn't a thing