grumpy got an update

This was about time. I built Grumpy several years ago with Debian 3.0 Sarge, and left apt sources pointed to stable (and testing for some time, as well as debian-multimedia, which sometimes has more recent versions of system software ...) I never ran apt-get dist-upgrade, and only updated packages I really needed, since it took me some time to get this system working properly with MythTV. I started this when MythTV 0.19 was the latest and greatest, and I didn't want to fiddle a lot with a computer that serves mainly as an entertainment system. What is Grumpy today has changed hardware twice.

Over time this had the funny side-effect of an only partially updated system. Yesterday, I finally took the plunge and brought grumpy up to Debian 5.0 (Lenny), a current kernel, new Nvidia drivers, the latest MythTV 0.23, and so on.

I followed the usual approach of pointing my apt sources at lenny explicitly, aptitude update and then aptitude safe-upgrade. This ran into an issue with /usr/bin/add-shell, which probably moved in etch from package passwd to package debian-utils, and passwd has a dependency on debian-utils. Some manual dpkg -i while ignoring dependencies got debian-utils installed, and then passwd installed properly. Remembering my days fighting with RPMs on RedHat and Fedora systems, I continue to be impressed by how well the DEB package installation process works. dpkg --configure -a became a close friend. Once that completed I ran aptitude full-upgrade and got to choose which dependency resolution approach I wanted to use. There weren't too many problems, mostly related to mythweb, apache, and kernel 2.6.18 dependencies. I pinned apache and removed some packages (I really didn't need gcc4.1) to get through this.

Afterwards I still needed to manually install the 2.6.26 kernel image, and the mktemp package (that was surprising).

Shortly before booting, I checked the raid status and noticed that the 1TB RAID partition holding all downloaded TV episodes we record in Germany, as well as my vacation videos was degraded. Sure enough, Nagios is checking disk full, but not raid status. Gaaah!

A little while later that Western Digital 1TB drive was replaced with a Seagate 2TB drive, and the partition was syncing from the partner drive. Then I broke the mirror the other way, and put in a second 2TB drive, since we were running somewhat low on disk space anyways. Videos and video editing eat up disk space very quickly. Nagios is using nrpe and check_md_raid (from exchange.nagios.org) to verify RAID health.

This reminded me again that I need to revamp the Nagios configuration. I set this up in a kinda crooked way years ago, and it didn't bother to make this more logical and straight-forward since. Maybe some other day...