Tales From The Geek Side

I’ve almost completed my first Ebay sale.  The auction has ended, I’ve received payment, and I’ve shipped the item.  Due to the holiday season the item has not arrived at the buyer’s address yet.  Overall I must say that I’m pretty impressed with how easy Ebay makes the whole process.  I currently have 4 auctions in progress.  I’m selling a printer, a SCSI controller, a bluetooth headset, and a sound card.  All of the auctions except the bluetooth headset are bundles — I’m selling a bundle of related items instead of single items.  So far I’ve got bids on everything except the headset.  I’m kind of hoping that auction ends with no sales.  Instead of selling it I could give it to a family member.  Had I been thinking I would have done that in the first place.

Tonight I put something for sale on ebay for the very first time.  Yeah, I know I’m a newb.  I’m too embarassed to say how long I spent crafting my listing.  The listing isn’t even for an expensive item.  I’m using a less expensive items to get my feet wet and then I will sell one of my guitars.

I must say that I’m disappointed with Ebay.  Their picture upload tool did not work with Firefox in Linux.  Come on!  Uploading pictures should not be browsers specific. 

I administer a couple of Debian based email servers all with similar setups. Each runs Exim, ClamAV, and Spamassassin. Around the time I upgraded the servers from sarge to etch I noticed that exim was logging to the panic log periodically. This hasn’t been a huge concern to me as it should result in a temporary delivery error and the sending email server should resend the failed message. Still any issue annoys me. I didn’t get very far tracking the problem down. As best I could tell ClamAV was taking longer than 3 minutes to reload virus definitions and was unresponsive to Exim during the reload period. Exim would then panic.

The log file contained entries that looked like this:

Recently a friend who seemed to have a similar problem alerted me to the Debian Volatile project. This project is, essentially, a backport project for software that is “volatile.” This includes software such as ClamAV and Spamassassin. It can be critical to keep up to date with these packages if you are running a mail server. A few days ago I upgraded Spamassassin and ClamAV to the versions provided in etch/volatile. This seems to have corrected the ClamAV problem I was faced and now I have a more recent version of Spamassassin to better fight spam with.

The problem it seems was that freshclam, the tool that updates the virus definitions, wasn’t able to get new definitions because it was too old. Apparently freshclam sends along version information and the server rejects the request if the version is too old.

I don’t update this blog often enough.  I should because it’s really not that hard to write a post (Wordpress is excellent) but it still seems to be inconvenient enough that I don’t post when I have something worthwhile to say.  I’m not saying that I have something worthwhile to say very often (this sentence is proof of itself). 

This week I came across a post on digg showcasing a number of blog editors.  I had looked into off-line blog editors on Linux and Windows not too long ago but I was not very impressed with what I saw.  This time I saw ScribeFire which is a firefox add-on.

I’m not usually wild about software that is implemented as add-ons or plugins for other software but in this case I think I’m ok with it.  Blogging is web oriented so embedding a blog editor in firefox could be justified.  Still I think I’d prefer it to be a standalone app especially since ScribeFire has add-ons of its own as well as themes. 

Time will tell if I continue to use this software or not but it looks very promising.

Today I installed WebSVN 2.0. I’m pretty impressed. It’s a very fast and useful web based tool for browsing subversion repositories. It’s so fast that I’m questioning why Tortoise is as slow as it is.

This is the first version that I was able to install due to security reasons. Previous versions effectively bypassed authorization settings because WebSVN would run as the web server user account. This version of WebSVN reads and honors authz files. All you need to do is configure Apache to authenticate users the same way as you would for SVN and tell WebSVN where your authz file is. This version also has a lot of usability improvements (no more tedious scrolling!) as well as a much improved look.

The stuff I like the most (so far):

• New look
• Honors authorization file
• No more scrolling when digging in a deep directory tree!
• Web based blame
• Web based diff
• Syntax highlighting via enscript

Today I paid my $15 to Schedules Direct.  That gives me three months access to their TV listing data which I use for MythTV.  The offer a free 1 week trial that I signed up for not too long ago.  When I paid my $15 via credit card through paypal I was surprised to see that my listing service was extended by three months.  I assumed they’d truncate the remainder of my trial period but they did not.  Their site reports that they have reached their break even point and it seems likely that they will be able to reduce the cost of the service to $20 per year.

Today I setup a trial account with Schedules Direct and configured my MythTV backend to use my lineups from Schedules Direct.  It went very smoothly.  The steps were:

• Create an account with Schedules Direct.  They offer free 1 week trials.  After a few days I’ll pay for 3 months.
• Update backend to latest from debian-multimedia.org
• Run mythtv-setup to change channel lineup source
• Run mythfilldatabase to grab listing information
• Test

Note that I did not have to update my myth frontends!
I had no hiccups at all along the way.  I was pleasantly surprised to find a nice simple interface for creating and managing lineups at schedules direct.  The only thing I’d like to see is a way to change the names of lineups.  I have an HD tuner card and an analog tuner card.  I like having a digital lineup and an analog lineup — both lineups are broadcasts.  To do this I have to use two different zip codes.  The result is that the names of the lineups are almost identical though the zip code is added to the name.  However, in mythtv-setup I can’t SEE the zip code (it’s truncated).  That makes setting the lineups a guessing game.

On most of my Internet facing Linux systems I have /tmp as a separate file system. I do this primarily so that I can set noexec and nosuid on the file system. While that won’t stop a real intruder it will slow down or stop script kiddies and worms.

This causes a minor issue with Debian. When I upgrade packages I’ve noticed that many packages create scripts in /tmp and try to execute them. That, of course, fails with noexec set on the file system. For a while I’ve been forcing myself to remember to remount /tmp prior to upgrading. This is error prone and I knew there must be a better way. Today I took the time to find a better way.

Apt is highly configurable but you’d never know it because you almost never need to changes default settings — at least that’s been my experience. Browsing through /usr/share/doc/apt/examples/configure-index.gz I found what I was looking for. You can specify shell code to run before and after apt invokes dpkg. I created a file /etc/apt/apt.conf.d/99local where I placed the following:

// Auto re-mounting of /tmp
Pre-Invoke {"mount -o remount,exec,suid /tmp";};
Post-Invoke {"mount -o remount,noexec,nosuid /tmp";};

Now I don’t have to remember to remount anything. This is also useful if you wanted to keep things like /usr/bin mounted readonly.

In the fall of 2005 a server that I operated was broken into. The perpetrators defaced every index.* file on the system including those in the archives of a mailing list that I’ve run since December of 2005. Recently it came to my attention that those files were overlooked in the cleanup process. I checked my backup CDs from 2005 and found that I did not have any backups that could be used to restore the defaced archives. So I did what anyone would do, I searched for ways to regenerate the archive. Mailman includes tools to regenerate archives but there are some issues.

Read the rest of this entry »

Earlier this year a friend of mine asked if he could give my email address and/or resume to recruiters at Google. Google set aside half of a day for engineers to look for resumes and contact friends who might be the type of person good would like to hire. I’d been happy with my current job so I was hesitant but I relented and said he could give them my email address. Over the next 3 months I experienced a level of disorganization that I never would have expected from a company with the Google’s reputation. The experience has soured my opinion of the company and it has given me insights on what to avoid when I take on the role of interviewer.

If you happen to be reading this and you work with me at my current job don’t get too excited. I told my manager that I was interviewing with Google. I didn’t want him to hear it from somewhere else and get the wrong idea. I was completely honest when I told him that there was very little chance I’d accept an offer if given one.

Read the rest of this entry »