It was actually a lot easier than I thought it might be to add encryption to a server's backup process.
I created a PGP GnuPG key, put the public key on the server, added it to the backup user's keyring, set the trust level, and added: gpg -r serveradmin@example.com -e backuptoencrypt.tgz && rm backuptoencrypt.tgz
Now the admins are the only ones with the private key, and anyone who gets ahold of the backups won't be able to do anything with them. Time to add this to more servers.
[[!tags backups encryption security gnupg pgp]]
The message queue for DBUS's org.freedesktop.Notifications filled up while I was testing it with notify-send. Evidently, it has a limit of 20 messages in the queue, and if the notification daemon happens to die and not remove itself from DBUS, the messages queue and aren't removed.
After looking into it, the notification interface doesn't have a way to clear the queue unless you happen to know the ID of the message. It doesn't even have a method to list current items in the queue!
I'm hoping that whenever I reboot the machine, it will clear that queue, but until then, I'm using the gnome-osd-client.
IMAPS is working again at work. I've been told that Exchange requires a mail-store restart when a cert is changed. I'm just glad to have access back.
I was pretty close to just forwarding all my mail to a zimbra server. Dealing with OWA/Outlook's limited email filtering was a real problem. I ended up sorting most of it by hand. Not fun.
Looking forward to the day we switch to a F/OSS email system, but I'm not holding out a lot of hope. The edu pricing on Exchange, and the resistance to change makes it hard.
We upgraded from PostgreSQL 7.4 to 9.0 at work over the winter break. I was much easier than the tests we did moving from 7.4 to 8.x Everything went fairly smoothly. The only trick has been with dblinks.
Evidently, PostgreSQL changed how they work, and we had to modify some things to get it working.
It's nice to be mostly rid of the unupgradable P3 733MHz single processor box that the 7.4 server was running on. Although, that box has been our main DB server for the website and the webapps we've written since before I started here in 2002!
That's a pretty good testament for Linux and PostgreSQL.
I use GIT for versioning projects I'm working on. I was able to use it to find a tricky bug today by hopping through commits.
It's pretty much the first time I've done that.
The bug was some PHP code that looks like this:
else
{
echo somefunction($var); } ?>
}
<td><?php
I looked at the HTML where the } were showing up, which was way at the top. (The jQuery DataTable additions pushed it even further away.) Which was way away from where it actually was. I thought it might be in the jQuery but had no luck after turning off a few of the included scripts.
I decided to track it down by hopping through commits, and found it after the second try. It feels good to have tools that help you fix things. Don't code without versioning!
I like the new Digital versions of Linux Journal. I constantly forget things so, it's nice to be able to read it anywhere on whichever device I remembered to bring, and not have to worry about carting around dead trees.
I know a lot of people aren't on the digital reading bandwagon, but I've been reading books on digital devices since I got my second PDA (a Handspring Visor Deluxe (the first was an old used Apple Newton M110)), and the sheer usefulness of being able to search and pull it up on any device outways the feel and smell of paper for me. I'll always have paper books (our current library has over 800 books), but I've also made sure I have digital versions, too. Which reminds me that I need to get back to that book scanning project.
I haven't noticed any decrease in quality of the articles or amount of content. I look forward to reading it for many years to come. Especially since I'm paid up for a few years. 
Speaking of Linux Journal. I always find Ruven Learner's articles to be useful and interesting. His articles are always about new technologies that I've heard about but haven't had the time to dig into yet. They're a great introduction to the topic.
His latest article is about ?wikipedia message queues. It got me thinking about how many different things there are on the back end that don't require holding up the client browser and that can really slow things down at scale. I'll start playing around with it and post anything interesting or what I end up doing.
We went to the Midwest Regional VMware Users Group meeting at Arrowhead stadium. Overall it was pretty good. They could have used more presentations by actual customers of implementing VMware setups. Handing out the give-aways was pretty buggy as there were a lot of people not present.
The keynote started off OK, but quickly slipped into a very detailed/technical view of the new version and put a bunch of people to sleep. He also went over by 1/2hr throwing off the schedule.
Arrowhead is a really nice location, and the catered food was really good, if pretty spicy.
Simple HTML DOM makes screenscraping a lot of fun. It's easy to use and does a good job of pulling things apart. Don't bother trying to regex (X)HTML. It will always hurt you.
Evidently, Cisco ASAs and AnyConnect VPNs don't like secure passwords. Passwords over 20 some odd characters fail at login. Fun. 
References
Updating old code can be a real pain if no organization is imposed either from the programmer, the programming language, or a framework. It's especially painful if it's your own code. PHP and Perl are extremely powerful languages, but without self discipline or some kind of framework, you're probably going to end up with sphagetti code.
I've been using ?CakePHP(http://cakephp.org/) and it is fantastic! It helps you with the complicated/hard stuff, and makes available components to autobuild code. It's in the middle of a line between trying to do everything for you, and making you do everything yourself.
You setup the database yourself and then point the cake scripts at it, and it can build a basic CRUD interface for you, or create a scaffold while you play with the database. I've used this to create a few QND interfaces in just a few minutes.
I upgraded my work desktop from ?Ubuntu(http://ubuntu.com/) 11.04 (Natty) to 11.10 (Oneiric). I had to manually disable the extra repos I've added in, and manually update the /etc/apt/source.list to oneiric.
The only problem I've had so far is that the fonts look like crap. Not sure what the deal is because it's fine on my laptop. Ah, yet another project for later.
I figured out the wordpress issue and updated the previous post. I was asking for it by taking a Debian testing setup and switching it back to stable.
This blog is powered by ikiwiki.