punkwalrus (punkwalrus) wrote,

Some Linux blues

I am still sick. Bleah. The sickness changes its methods over time, for instance right now, my fever is almost gone, but I have a lot of congestion. This is my personal “critical point” because what happens now may determine whether this stays a cold, or turns into pneumonia. Again. So I am drinking expectorant round the clock, as directed. And taking echinacea. And trying to stay warm. I am at work, because I get the pager again for the next two weeks, and I think I can make it until the end of today if I stay calm.

This weekend, I didn’t get a whole lot done. I balanced the checkbook, as I mentioned earlier, so my finances are pretty up-to-date. I cleaned the cat box. And I finally got around to upgrading my Linux box, which was more work than I would have liked.

The “hub” of all my home’s data is a central server called “pippi.” Pippi is a P4/3.0ghz HT box with a GB of RAM, a RAID5 array for all data, and single disk for root. Over a year ago, I think, it done blew up on me. Turns out the power supply sucked, and that was replaced, but when that happened, I lost some of my RAM, my motherboard, my video card, the root drive, and one hard drive in the RAID array. One rebuild later, it was on Linux, FC3. I had to replace the root hard drive with a 20gb spare I had lying around, got a new HD for the RAID, and used a spare, 7-year old crappy ATI Rage card for the video.

Time passed.

I kept putting off upgrading the box because I was lazy. Finally, after the Fedora Legacy project was scrapped, I was forced to contemplate the upgrade, and decided this weekend was as good as any. Was this simple? Oh no. It never is as simple as you think it’s going to be anyway, is it? Here’s my timeline:

1000: Tried to burn FC6 DVD. k3b (really, growisofs) told me I was using a double medium DVD, and 50% of the space would be wasted! I didn’t care. But it did! It wasn’t even that it couldn’t do it, it simply wouldn’t! Known bug.
1030: Didn’t have 6 CDs to burn for FC6, so I burned the first one, and wanted to do a network install. On boot, I did “linux askmethod” and chose FTP. Then it launched into Anaconda (GUI install) and crashed. Tried text install, and
11:00: Decided to do “hard drive install” because I had the ISOs on my USB backup drive. GUI hung at some xml issue. Went ahead with text install. After hanging for 10 minutes on the upgrade (again, a known bug), it went ahead and started the install, but then said I needed 571mb more space on the hard drive to continue. Freeing 1gb resulted in an error that said I needed 241 more. I gave up. I knew the 20gb drive was running low on space, and I needed to upgrade to a spare 40gb I had set aside, but that meant a flat new install. Sigh. Luckily, I had backed everything up on the USB drive the night before in case the upgrade went all wrong.
12:00: I took apart the Linux box, blew out some dust. Replaced video card, root hard drive and checked the cable connections. Blessed the box. Rebooted.
13:00: Started fresh install. Decided to do “linux text askmethod” to be safe. While selecting packages, it suddenly started to install. WTF? Oh well, I can always add/remove packages later. Let me tell you, doing an install off of images on the hard drive? Really fast. Like 12 minutes later, a full default gnome install was ready to go.
13:15: Rebuild array. While the installer detected a software RAID5, it did not assemble it. I also found out FC6 has a problem with detecting USB drives on boot. This seems to be fixed by turning off and turning on USB drive. Then it loads /dev/sda1, and I can mount it. Loaded mdadm.conf, and rebuilt array md0, which had /home. Added to fstab. Rebooted, and /home still came back. Awesome. Can’t write to it, though. Not awesome. Can’t write to USB drive either. A quick show of the logs says that there are SELinux kernel errors. FC6 loaded SELinux on default.
15:00: Fixed the r/w errors on md0. Not sure how. I think it got fixed in some reboot. Set up users, but predicting permissions problems, I made sure I restored exactly what their UID and GIDs were previously. Started working on samba shares. used smbpasswd to add users, and when that didn’t work, I discovered Samba was installed, but as a client only. Grrr... installed samba server...
17:00: After wrestling with samba shares, I finally removed SELinux from the system because it was just pouring kernel errors into the logs. This seemed to fix a lot of problems right away. I made a mental note to understand how SELinux works, because I am sure SELinux was not broken, just not configured properly. Samba didn’t work at first, but then rebooting the Windows box I was using fixed the problems. Samba shares seemed snappier. This may be because I haven’t run SETI on this box yet. Luckily, I set up SETI as a user with all the needed files in the home directory.
19:00: Finally got most of the box as I wanted for now. Monitoring needs set up again, Cacti/Nagios needs reinstalled, but I don’t want to mess with SQL right now.
Tags: fedora, linux, sick
Comments for this post were disabled by the author