BOINC has been added to the Fedora 7 and higher repositories. You can use your favorite GUI front end or just do

# yum install boinc-client boinc-manager

The cool part is my logrotate config file is in the repositories as part of the package. I don’t know how geeky that makes me, but it sure did make me feel good. I don’t program, so there’s isn’t a whole lot I can do to give back to the Fedora community. So I think it’s cool that I was able to contribute in this small way.

The capital letter in my user name was indeed throwing sendmail into fits. I could have created an entire new user. But hey, that’s a lot of work. I’d have to port my Firefox profile over and re-customize my desktop. Nah. Definitely not worth it just to be able to read local email.

I could have directly edited /etc/passwd using vipw at the command line. I’m not a big fan of vi. It scares the pants off me. I’ll stick to emacs, thanks. Heck, I find emacs extremely challenging on a good day. I’m all for the simple route of using kedit. So on the advice of mweltin who was hanging around #boinc, I ended up using usermod from the command line. The syntax was quite simple. usermod -l kathryn Kathryn. The only gotcha is that you can’t be logged into an account that you want to edit. So I logged out of Kathryn. Hmmm… well that dumped me to a text login screen. Oh well. I logged in as root, ran my usermod command and then rebooted from the command line. I have no idea how to get back into a graphical environment from a text prompt. But at least I learned how to reboot from the command line.

And…. BINGO!

[kathryn@Galaxy ~]$ mail
Mail version 8.1 6/6/93. Type ? for help.
“/var/spool/mail/kathryn”: 6 messages 6 new
>N 1 boinc@Galaxy.Fedora Mon Feb 11 19:42 17/582 “test”
N 2 MAILER-DAEMON@Galaxy Mon Feb 11 19:42 78/2716 “Warning: could not send message for past 4 hours”
N 3 MAILER-DAEMON@Galaxy Mon Feb 11 19:42 164/5257 “Warning: could not send message for past 4 hours”
N 4 root@Galaxy.Fedora Mon Feb 11 19:42 27/978 “Cron <root@Galaxy> run-parts /etc/cron.daily”
N 5 logwatch@Galaxy.Fedo Mon Feb 11 19:42 113/3519 “Logwatch for galaxy.fedora (Linux)”

I found an answer to my mail problems. Someone hit on my blog using sendmail “user unknown” errors 8.14.1 as a search term. I figured that this person was having problems similar to my own. So I put that same search in Google and found this thread.

It looks like having a user name with capital letters in it throws sendmail into a tizzy. This fits perfectly with what I observed on my own machine. I can receive mail sent to root or boinc, but I can’t receive mail sent to Kathryn.

So I guess I need to create a new user for myself called “kathryn”. I’m just not sure what to do with all the stuff that’s in “Kathryn”.

Oh well. At least the mystery has been solved.

I’m sure SELinux is a great thing. But at the moment, it’s getting in my way. So I found out from my wonderful teacher how to disable it. This is one of those notes to self type post…

Make damn sure you’ve edited the correct line in /etc/sysconfig/selinux. If you don’t, you get scary stuff like this when you reboot…

Unable to load SELinux Policy. Machine is in enforcing mode. Halting now.

Kernel panic – not syncing. Attempted to kill init!

Yes boys and girls, that freaked me out just a little bit. I swear I never tried to kill init. So what did I do wrong?

Here are the contents of the default /etc/sysconfig/selinx

# This file controls the state of SELinux on the system.

 # SELINUX= can take one of these three values:

 #    enforcing - SELinux security policy is enforced.

 #    permissive - SELinux prints warnings instead of enforcing.

 #    disabled - SELinux is fully disabled.

 SELINUX=enforcing

 # SELINUXTYPE= type of policy in use. Possible values are:

 #    targeted - Only targeted network daemons are protected.

 #    strict - Full SELinux protection.

 SELINUXTYPE=targeted

# SETLOCALDEFS= Check local definition changes

 SETLOCALDEFS=0

Here’s what I did on my first attempt.

# This file controls the state of SELinux on the system.

# SELINUX= can take one of these three values:

#    enforcing - SELinux security policy is enforced.

#    permissive - SELinux prints warnings instead of enforcing.

#    disabled - SELinux is fully disabled.

SELINUX=enforcing

# SELINUXTYPE= type of policy in use. Possible values are:

#    targeted - Only targeted network daemons are protected.

#    strict - Full SELinux protection.

SELINUXTYPE=disabled

# SETLOCALDEFS= Check local definition changesSETLOCALDEFS=0

And here’s what I should have done.

# This file controls the state of SELinux on the system.

 # SELINUX= can take one of these three values:

 #    enforcing - SELinux security policy is enforced.

 #    permissive - SELinux prints warnings instead of enforcing.

 #    disabled - SELinux is fully disabled.

 SELINUX=disabled

 # SELINUXTYPE= type of policy in use. Possible values are:

 #    targeted - Only targeted network daemons are protected.

 #    strict - Full SELinux protection.

 SELINUXTYPE=targeted

# SETLOCALDEFS= Check local definition changes

 SETLOCALDEFS=0

Notice that on my first attempt, I changed the wrong line. DOH!!!!

So how did I fix it? I’ve come to the conclusion that triple booting is a very good thing. Vista won’t help much in these situations because it can’t natively read, let alone write to an ext3 partition. However, having a second Linux distro installed can be very handy. I booted into Ubuntu and mounted my Fedora partition according to the notes I took a while back. It’s easy. You create a mount point by giving the command (as root) mkdir -p /mnt/sda6 (sda6 happens to be my Fedora partition) and then you mount it by doing mount /dev/sda6 /mnt/sda6. Once I had my Fedora partition mounted, I started gedit and opened /etc/sysconfig/selinux and changed what I needed to change. I saved. I rebooted. And bingo. I’m back in Fedora!

The one thing I haven’t written about here is the intense fear I had of anything to do with the terminal. Just thinking about it would send me into fits of panic. I was convinced that one mistyped command had the power to turn my computer into a small black hole. Try as I might, I really haven’t managed to break anything. And the things that did go wrong ended up being easily fixed. What would I have done if I didn’t have a Ubuntu partition? Booted from a LiveCD and done the same thing.

I’m in the process of testing out a rpm for BOINC. Woo hoo!

Some notes to self:

• After creating a tarball of one’s working install. One needs to move it somewhere for safe keeping. One’s data partition is a great place for said safe keeping. When one needs to move something to said partition, it helps if one mounts said partition before trying to move it. Somehow I think I now have a /Data directory hanging out somewhere. I’ll have to remove it once I find it.
• rmdir only works on empty directories. If there are files in there, you have to use rm -r.
• Using rm -r on a directory that has a metric ton of files takes a very long time because it runs in interactive mode. If one is really sure one wants to get rid of the whole thing, use rm -rf.
• Once you remove a directory containing a file, the soft links pointing to said file break. And programs that need said link to file to function correctly will also break.
• It is impossible to echo something > broken_soft_link_name. You must rm broken_soft_link_name first.
• Wiping out a users home directory is a Bad Thing ®. Not having all the hidden config files causes Very Bad Things ® to happen, including not being able to access the account from the shell. Having a backup is a Good Thing ®. Knowing how to restore said backup is an Even Better Thing ®. And finding out that said restored backup works is an Extremely Good Thing ®.

.

Watch this space for further updates…

It’s no secret I’m actively involved with BOINC. I’ve learned a lot about computers from running their software and helping out in the various project forums. It was one of the project administrators that so kindly helped me get Linux up and running on my computer (current uptime 15 days, 10:07 and still going strong). Because running BOINC tends to stress your system, I was looking for a lightweight monitor for things like memory usage and temperatures. I know all about top to get processor/memory usage, but having nice little program with everything in the same place is preferable. So I found Conky.

If I had Debian or Gentoo or just about any other Linux distro, I would have been free and clear because there are pre-compiled packages. But of course there wasn’t an rpm for it (I use Fedora 7).

Compiling from source was supposed to be as easy as configure, make, make install. Note the use of the term “was supposed to be”. I guess if I actually had my system set up as a development box (and I actually knew slightly more than nothing) it would have been easy. My experience with compiling from source was limited to mplayer and that basically just worked (once I got g++ installed).

I knew I was going to need the X11 development libraries so I installed those. I bravely opened up the terminal (actually I’ve gotten quite comfortable with it) and unpacked the source archive. I changed over to that directory and told it to go do the configure. I watched it churn away and spit out an error. Grumble grumble grumble. This was a few weeks ago. I asked a friend for help, but he wasn’t familiar with the libs. I Googled. No joy. Nothing looked relevant and what I read went so far over my head that I needed binoculars to see it. I posted on the Conky forums, but no response. So I put the whole project on the back burner for a few weeks.

I was bored this morning and decided to see if anyone on the BOINC forums might know where I needed to go from where I was. Within thirty minutes I had a response. Thank you Trog Dog! Just by seeing what I needed to install I also saw how to read the error message. So I installed what I needed and ran configure again. Another error. So I downloaded and installed some more developer stuff. Later, rinse, repeat about three more times. And then the configure succeeded. Yippee!

Make and install went flawlessly. I ran the program. And BINGO! I now have Conky on my desktop.

Now I have some reading to do on how to best configure Conky for my use. What is shown in the picture is the default configuration. But after getting it compiled I needed to go and finish my grocery shopping. So that little project will have to wait for another day.

I wasn’t happy with just having my laptop dual booting between Ubuntu and Vista. Nah. I had given up on the XP/Vista/Ubuntu triple boot idea mostly because the XP Pro CD I have doesn’t have the right drivers for my SCSI hard drive.

In working through the terminal stuff with Eric, he sort of turned me onto the idea of trying Fedora Core. So we came up with a partitioning scheme that will allow me to triple boot among Vista, Ubuntu and Fedora Core 7.

The partitioning was the easy part. Once I remembered that I had to reboot after wiping out the last three of the four partitions I already had setup, it went flawlessly.

For the curious, my partition table looks like

/dev/sda1 — Vista

/dev/sda2 — Shared data

/dev/sda3 — Swap

/dev/sda4 — An extended partition containing…

/dev/sda5 — Ubuntu 7.04

/dev/sda6 — Fedora Core 7

/dev/sda7 — /home

/dev/sda8 — /tmp

So now I’m at the point where I’m ready to scream. My first attempt was a failure. I didn’t pay attention to the GRUB (bootloader) screens when I installed Fedora so I ended up not being able to boot into Ubuntu. Thankfully I was still able to get into Windows.

After consulting with Eric, I decided the easiest thing to do was reinstall Fedora and add Ubuntu into GRUB. So I did that.

Failure again. Although I now have the option of booting into Ubuntu, I still can’t because it throws an error at me. But again, at least I can still get into Windows.

So now I’m off to try to mount my Ubuntu partition in Fedora through some CLI magic (and it really is magic to me) and then find the names of vmlinux and initrd (whatever those might be).