Archive for August, 2009

Site Codes, an Inside guide to Atlantic Metro’s Network Locations

Monday, August 31st, 2009

What are site codes ?
In most organizations we highly recommend the use of site codes as even single-site companies often ends up with remote offices, disaster recovery sites, or additional data center locations over time.

The goal with site codes is to choose an identifier that people both from the site in question and other far away can easily identify as being related to a given location. We have struggled with this in the early stages of growing our business, as there was lots of potential for confusion and overlap.

Choosing a Site Code Convention
There are a few ways to go about choosing a standard for site codes, and over the years we have found its best to use Airport codes.

Example – Newark, NJ | EWR (airport code)

In-depth look into Atlantic Metro’s Sites

LGA0 – TELEHOUSE – 25 Broadway

LGA1 – Atlantic Metro IDC – 325 Hudson st

LGA2 – Telx – 60 Hudson st

LGA3 – Switch & Data – 76 9th ave / 111 8th ave

LGA6 – CORESITE – 32 6th ave

LGA7 – TELX – 111 8th ave

NJ3 – Atlantic Metro IDC – 5851 West Side ave

ORD0 – Atlantic Metro IDC – 350 E Cermak

ORD1 – Equinix – 350 E Cermak

ORD2 – CORESITE – 427 S. LaSalle

LAX0 – CORESITE – 624 South Grand Ave

YYZ0 – TELEHOUSE – 151 Front St.

Don’t be lazy, make your own local yum repo…

Wednesday, August 19th, 2009

So I often log into CentOS/RHEL servers and discover that there are tons of custom RPMs installed that were just downloaded from random places. While I understand that sometimes you just need to get things done, please just create your yum repo somewhere and install from there. You’ll thank me later when you need to update said packages on 20+ servers.

Here’s the basic gist.

On what you plan to be your yum repo server:

1. Copy all your custom/cool RPMs over to some web accessible directory on a server. (optional: maybe create some cool vhost in Apache or whatever to have a name like imnotlazyandmakemyownyumrepos.mydomain.com)

2. Make sure yum-utils is installed (available via default Cent/RHEL repost) “yum install yum-utils”

3. Use the createrepo command to do all the cool yum directory structure magic and such “createrepo /var/www/my_yum_packages”

On your yum clients:

1. Subscribe to your new/cool repo by creating a file in /etc/yum.repos.d/ “vi /etc/yum.repos.d/cooldude.repo” (must end in .repo)

2. That file should have contents like this (note, you can do all gpg/arch stuff if you want but my repo is secured internal and it automatically figures out which arch works for what machines, so I don’t see the need currently):


[MY COOL REPO]
name=My Cool Repo
baseurl=http://imnotlazyandmakemyownyumrepos.mydomain.com/my_yum_packages
enabled=1
gpgcheck=0

3. You should be all set to use your own cool yum packages. Try a search to verify it worked “yum search mycoolpackage”

4. If you don’t see your cool new stuff, yum’s caching mechanism is sometimes a tad overzelous. Go ahead and run “yum clean all” to get a nice fresh start.

Next time you need to update php to your cool/custom build on like 20 servers, just dump it your repo server, re-run the creatrepo command, and do a “yum update php” on all your servers. If there are new dependencies required, just throw a copy of it in your repo as well.

Much better, right?
Need help with making your own Yum repo ? Atlantic Metro offers system administration services, contact us today – sales@atlanticmetro.net


is Digg proof thanks to caching by WP Super Cache!