Muh Ha Ha Ha! Finally, I think, everything is back online. To resolve the missing database (thus no blog) issue, several other issues, and to give me a little sense of security. I migrated all of the key systems and files from a small(ish) RAID1 array to a much larger RAID5 array over the weekend. And, at the same time I upgraded from Fedora Core 4 to 5. While the upgrades went suprisingly well it has taken several days to get everything back online (and configured as they were). Now the server can get back to doing all it’s old jobs as well as some new, high importance tasks - like database replication…

The list of issues with eMail was getting large, and it’s load on our system was getting bigger and bigger. I finally bit the bullet… It wasn’t pretty (Kel was pretty tense for a couple of days) but things are now back up and running better then ever!

Shortly after 10:00 am Verizon arrived for their scheduled 4 hour FIOS installation - not too bad considering I was told they could arrive anytime between 8:00 am and 5:00 pm! The first thing the installer said was, “There’s no conduit here…” I was immediately overcome with a wave of sadness. Until moments later a huge truck dragging a Ditch Witch, a monsterous spool of conduit and a team on non-English speaking shovel toting conduit installers arrived!

I’ve put up with Comcast’s lousy service for almost 2 years now - because I haven’t had any alternatives. I pay $57.95 / month for 6 Mbps / 768 Kbps and I’m lucky if I get 1 Mbps / 384 Kbps. But tommorrow, that is all about to change! If all goes according to plan all of the bright colored paint across my driveway and yard will be dug up and Verizon will lay fiber from the street, up the hill that is my yard, to my house! With FIOS I should be getting 15 Mbps / 2 Mbps for $44.95 / month. $13 cheaper / month for more then twice Comcast’s published speeds, but realisitically 15 times their actual speeds. It’s been an awful long time coming, but I’m sure it’ll have been worth the wait.

Dang it’s been a long time! Curling has wrapped up - I had a disappointing year, whereas Meg had a repeat championship victory! Softball has started - the JoanneHudson.com Cubs & Red Socks are busy practicing. So are the Quix Stix Waves, for that matter. Spring Break is rapidly approaching and we’re changing our plans like the weather… But, the biggest news of all… Are you ready for this? I JUST SCHEDULED MY FIOS INSTALLATION! WOO HOO! It’s almost too good to be true, but according to the Verizon website the fiber that was laid last August is now bright and our humble abode qualifies! I have a feeling that I’m going to be crushed, but my scheduled install date is April 5th none the less.

There is something seriously wrong with MovableType (the underlying software for this blog). Any kind of adminstrative action (posting, config, managing comments, etc.) takes forever, can only be done from the server, and is essentially a crap shoot (if it’ll work or not). There are mt.cgi processes dieing (going defunct) like crazy too. Jon says that he is having similar problems with MT on OSX too. A quick search of the Six Apart support forums show that there was a similar issue with MT back in 2003, but there’s nobody (other then me) talking about the problem now. I’m at a loss as to what’s going on, but I am aware that there’s a problem.

With little deviation from the common goal I’ve always supported the evil Microsoft empire - like them or not. I ran Windows386, and put up with all it’s crashes and memory issues, when everyone else was still running DOS 5.0. My only real divergence from the master plan came after Windows386; frustrated by it’s instability I moved to OS/2 (2.0 through Warp). But even then I guess I was still running Microsoft, as they were in bed with IBM and Windows 3.0 (and later 3.1) were dynamically linked.

I’ve been bitching lately (why should today be any different then any other day) about the cost of electricity. It seems that we keep using more and more of it - last month we used over 2800 KWH. So it was with great enthusiasm when Jon offered me a Kill A Watt electrical usage monitoring device. I brought the device home and put it right on the server and discovered that it was using just over 400 watts - so, in essence it’s costing me about $23 / month to run it. That’ll probably be one of my big hitters, but I still have 2400 KWH to account for - I can tell I’m going to get some real use out of this device.

Now that I’m running Linksys’ unreleased firmware (1.3.4) on my RV0041 it seems like all my VPN issues have vanished. So what better time to make a change to things! I decided that I was going to replace my WAN backbone with fiber, and leverage the RV0041’s built in mini-gbic sfp socket - but I didn’t get very far. It seems that Linksys makes 3 mini-gbic transceivers (2 fiber, 1 cat5) and none of them are compatible with the RV0041. Further investigation reveals that there is, in fact, only 1 sfp that will work with the RV0041: Copax’ LPS-301F2-XX. Unfortunately, it also appears that Copax’ products are only sold in Taiwan… Oh well, so much for fiber.

After several weeks of trouble shooting, I think I’ve isolated my network problems. To recap / bring you up to speed: My 8 Mbps connection peaks at about 6 Mbps, then gradually drops to under 1 Mbps. My VPN tunnels stay nailed up for about an hour and then all fail simultaneously - attempts to reconnect them always fail, unless my Linksys / Cisco RV0041 router is IPLd. To try and resolve the problem I’ve moved my cable modem as close to the network demarkation point as possible (2”). I’ve replaced the cable modem. I’ve removed all extraneous switches from my network. And I’ve run fiber from the modem to the router. So last night, when my VPN tunnels failed, and I went to restart the router (from its system management screen) the router woudn’t let me in I started looking at that piece of hardware…