Matt McCarrell - MP451 (January 2010 - May 2010)

Animal

Rebuilt twice this semester due to hardware failures on the IBM storage units.
Backups are working normally again. (Backups pushed every night from production systems to Animal via a script that pipes tar over ssh. Those backups are then pushed from Animal to PLBackup1 in the projects lab at 12 PM daily.)
Appears to be stable now and I haven't seen any failures in the weekly consistency checks since the last rebuild.

I've been fixing space issues as necessary. No longer a problem since we stopped mirroring freebsd.
Added Unity Linux and Moblin
Updated site template to use COSI design
Hoping to fix ownership issues for the content we mirror. We currently run our rsync scripts as root and the owner/group of those directories/files is whatever they are on the master servers we sync from.
I updated some of our upstream mirrors to sync from official location since we were given extra outbound bandwidth earlier in the semester.
Hoping to convince OIT to give us more outbound bandwidth one the semester ends.
Let me know if you want to mirror something. Over the summer is a good time for us to pull new content since the campus has bandwidth to spare. The new Ubuntu beta is currently there and the final release should be available around the release date

Replaced our old DNS server with two new servers (dns1 & dns2).
No longer hosting an open DNS server... wooops!
New DNS servers are running BIND and are in a master/slave configuration. This allows *.cslabs.clarkson.edu to still resolve in the event that one of the two servers is down.
Each of the new DNS servers is hosted on different dom0s to provide extra redundancy.
Most of our servers and some of the lab PCs are now using our DNS servers. This allows us to avoid potential netreg issues and gives us the flexibility to modify our upstream DNS resolvers to something like OpenDNS.
Systems seem fairly stable and have been up for 72 days without any issues.

Working on a new redundant LDAP setup using CentOS Directory Server
Currently have part of the dev setup complete but I need to redeploy it with some sample ldap entries to verify that my configuration is correct.
I'm going to test this more over the summer and hopefully integrate it in with our current setup before next semester.

Xen1, Xen2, and Righteous appear stable and only minor changes have been made to them. Xen2 has not crashed in over 200 hundred days. I'm not quite sure if a kernel update fixed this or if it's due to the load on the system being fairly low.
Trevor and Ryan assembled Xen3 earlier in the semester.
Trevor and I racked and cabled the system. We also deployed CentOS & Xen on the box.
I've been working with Trevor to teach him how to create VMs and use Xen. I will continue to teach him over the summer and next year.
I have created a few new VMs as needed using my generic VM script and have been updating the Ubuntu generic VM on a regular basis. I'm also planning to create a new Ubuntu generic VM based on the new 10.04 release once it comes out next week.
I worked on the CentOS generic VM a bit earlier in the semester but I need to finish configuring the VM and touching up the script. I'm hoping to finish this over the summer.

I have not managed to make any progress on the storage server sadly. I am planning to get an OS on it prior to leaving for the summer, which will allow progress to be made on it remotely.

Convinced Ryan to patch the VM earlier in the semester. The VM didn't have the Debian security repo added and thus had quite a few patches available.
Discovered that over the past few weeks the load on Docs and the CPU usage was higher than normal. Ryan fixed this by blocking the msnbot from crawling the site.
Attempting to keep content updated.
Deleted a lot of old pages to avoid them confusing people and because many of them had little to no information.
My contributions to the wiki can be found here.
My Nagios tutorial still gets viewed several hundred times a month and currently has over 9000 views.

I've added new systems to Nagios, AWStats, & Cacti as needed. If you would like access to any of these (since there is a lot of interesting info.), let me know.
I'm working with Bryan Rose to add more components to our monitoring. I believe he is currently investigating Snort for use in the labs.
I finally documented the Cacti install on Netstat.
I created a generic set of instructions for setting up Cacti on CentOS 5.

Converted Roger's design to an easy to use template. (This page is using that design!)
Rolled out COSI site design to almost every site except some project sites and the ITL site. I'm hoping to get the design rolled out to the remaining sites at some point.
Added a lot of our public sites to Google webmaster tools. This has helped boost our page ranking and allows us to detect for malware as well.

I have been trying to post whenever I'm working on something. I've found this useful to provide news pages for systems like mirror. All of my posts can be viewed here.
I updated the install earlier in the semester.
I have been pushing for more people to use it and some of them are!
I transferred ownership of this project to Trevor. He upgraded us to the latest release... after a few failed attempts and me having to restore it from backup a few times. ;)

I have been patching the systems I maintain on a regular basis. Currently, all systems I maintain are fully patched.
Disabled unneeded services and made several other changes (see my wiki changes).
Installed ESXi on most of the blades to provide an easy way to create VMs that may need to run Windows or be temporary for various testing.
I am working with Trevor to teach him about each of our systems and will be delegating tasks to him over the summer and next year.
Tonight, we will have 300 days of uptime for docs, mirror, mysql, netstat, righteous, and web1. Uptimes for several other systems are around 230 days and plbackup1 is around 290 days. I believe this may be the longest we've had a system up for since I've been at CU (*knocks on wood*).