Upcoming Downtime
Gollum (this server) will be under going some emergency maintainance later today (Sunday about 3PM) to replace a failing processor fan. We’ll hopefully be down no longer than an hour, but you know how it goes …
Gollum (this server) will be under going some emergency maintainance later today (Sunday about 3PM) to replace a failing processor fan. We’ll hopefully be down no longer than an hour, but you know how it goes …
Well, at 3AM CDT on Sunday morning (9th) the webserver went down. If you noticed it, thanks a ton for not reporting it ;-) I’m really surprised no one notified me as I didn’t notice it until about 4PM. It turns out that it was all related to an upgrade of Apache I performed on Friday, but didn’t manifest itself until Apache restarted Sunday morning as part of some weekly auto-cleanup scripts I have in place.
Gentoo changed the way that the Apache configurations (and modules, and …) are stored so as to be more mainstream. This seriously broke my custom configuration and effectively knocked all my websites offline. To be fair, Gentoo did announce that these changes were happening, I just forgot that it was this version they were doing it to and didn’t watch the compile so I missed the notice at the end warning me about this problem.
It took me about an hour to get the config all straightened out and modules rebuilt, but in the end I think it was worth the time. If there’s a website that I host that you cannot currently load, please let me know immediately so I can take a look.
OK, so things worked out great this morning. The cable tech showed up early of all things. He made some changes at the distribution point for the apartment complex and then stopped by to check on things inside. It turns out that the complex has four coaxial lines coming from the cable company that serve as the distro for the whole complex. Apparently the tech moved my service over onto one all by itself. I don’t share it with anyone now! This earned a 4db gain in signal strength, which equates to no more connection problems. He did say that my upload strength is close to borderline, but it appears to be working ok right now. I haven’t tested with the TV and modem yet, but if that doesn’t work, there isn’t much that can be done without inspecting the cable from the distro point into my apartment. It’s possible that there is some ungodly amount of cable spooled up somewhere that is causing the drop in signal strength to me.
I’m having nothing but problems right now with my cable connection, so if you can actually see this, rejoice as it will likely go back down soon. It appears the problem is signal strength yet again as the modem won’t even connect most of the time, and when it does, don’t even think about hooking up the TV as that causes immediate disconnection.
I called the office (Longview Cable) this morning to get this resolved. The lady I talked to was very nice and sounded like she really wanted to help me get this taken care of once and for all. I was really upset by the time I called, but I tried to remain as absolutely civil as I could and I’m glad I did as she probably wouldn’t have dealt as well with me had I started out by biting off her head :)
She got me on their schedule for tomorrow morning (10 - 12), so I will be at my apartment during that time span waiting for a tech to show up. I really hope this tech doesn’t give me a problem. I probably know more about networking (and even cable modems) than most of the techs I’ve dealt with up here. I know what the problem is, I just hope I can convince this person quickly enough to actually test it.
Frustration bleeding off … will just have to wait until tomorrow.
OK, web-based email seems to be working. I don’t know what caused the bottleneck last night, but it seems to have cleared overnight. To complicate matters, Gmail seems to be responding very slow the last few days and I think that messed with my tests.
OK, I admit it … I broke the email server. We are recieving email just fine, but it does not seem to be sending email at all … even from localhost, which is a bit scary. I’ll be making some backups of all our email in the morning (not that it is at risk) just in case I screw something up royally while trying to get it to work. I still have my main config files from before the upgrade, so I might be able to drop back to the last working state if it looks like it will be a while. Stay tuned for more information.
I will be doing some planned maintainance over the next few days. It’s well past time to upgrade some of the installed versions of the email server software and I’m not sure how smooth that process will go. Hopefully it will go well, but for anyone who might look here first … be forewarned.
Not my fault! I hate CableLynx. At least the tech guy that answered the first time seemed nice and relatively intelligent. Apparently I was one of the first people to call about the outage and as a result, they didn’t have a whole lot of information for me at the time. That was at 7:00PM … things just came back up a few minutes ago, so about 8:30PM.
This time, it is scheduled. I will be shutting down Gollum for a couple of minutes tonight about 21:00 CST. It’s time to add some much needed RAM to the machine. I’ve been holding off because I like my uptime stats, but since I had to shut down Sunday due to the power outage I have nothing to lose. If all goes well, Gollum will be down only long enough for me to drop a new stick of RAM in and reboot. After that, we should be at 1GB of RAM instead of the current 512MB (which seems to stay pretty full these days).
UPDATE: Hardware installation complete as of 21:21 CST. Gollum now has 1GB of RAM and a new network card (which I almost forgot to install until I saw it stored with the RAM). If anyone notices weird behavior, let me know.
This time it wasn’t my fault. I had a power outage here at the apartment that caused me to shut down Gollum. Of course the server was offline even before then because my router and cable modem aren’t on the UPS (sounds like something to fix this afternoon). The UPS held up suprisingly well, although I manually shut the machine down after about 30 minutes just because I wasn’t sure how much longer we were going to be without power (as it turned out, about another 30 minutes). Anyways, it was out of commission from about 15:10 - 16:30 CST.