Older blog entries for Mulad (starting at number 17)

Entry 17

Well, I decided that I'm making enough money to go out and buy a new large IDE hard drive. I may regret this. Oh well, I can always sell something, like my Palm II, which has been going fairly unused..

Now I'm stuck with the task of figuring out what exactly to do with 60 GB of disk space. My /:/usr disk is 8 GB, and I figure I can move my 2 GB /home onto that disk. Of course, my big issue is that I want a decent filesystem now. I'm not sure how long an e2fsck will take on that much space, and I don't think I want to find out. I mentioned earlier that I see ReiserFS, Ext3, and GFS as options. I worry a bit about Ext3, even though I haven't tried it. Ext2 (semi-)compatility is nice.. I wonder if either ReiserFS or GFS have anything like Ext2's file attributes: immutable, append-only, etc. However, both ReiserFS and GFS are supposed to be significantly faster at many things than Ext2(/3?).

One of my concerns is that of large file support. Not that I expect to be coming across many large files (>2 GB). Apparently, this will all become moot when 2.4.x arrives, although GFS does support large files in 2.2.x. The other thing is journaling. Ext3 is basically designed to add journaling to Ext2. ReiserFS has had journaling for over a year. GFS's journaling is relatively new, added since their last release in late 1999 (it's in CVS). GFS really scores points for being a best-of-breed clustering filesystem (but I'll only be running it as a local FS, so that doesn't really matter), though Reiser counters by having interesting ideas behind it like behaving almost more like a database than a filesystem. If I had an Alpha, I'd probably jump for GFS. It's largely been developed on Alphas, where Reiser is geared more toward i386.

I understand that there were people at OLS from all of these different camps. The GFS people have some slides..

Blah, enough about filesystems. Hmm.. Oops! I forgot about a backup that I was supposed to be doing.. I'd better go do that.

Entry 16

I'm finally learning a bit more about why I'm here doing this job that I'm doing. The business school where I work is trying to set up a fairly massive Oracle database which will contain all sorts of information, from who is enrolled in which class to the financial data that we have in a number of binary or ASCII data files on our main servers. It's sounding like it will be a massive thing. We've got a quad-processor Sun box (E425R, I think) with a RAID bank that will be running the database, eventually accessible via web, Java, and other front-ends (or so I'm told). Request Brokers and all of those fancy buzzwords were thrown at me. Certainly, Oracle is a nice database, but can't you just slap a PHP frontend on it and call it a day? (At least in the early stages?) Also, it seems like my efforts around here will go to waste if people are no longer logging into my boxen.. A web interface is just so sterile and with zero personality.. Then again, Solaris' ordinary personality really, really sucks..

Last login: Fri Aug  4 11:16:11 2000 from <...>
Sun Microsystems Inc.   SunOS 5.7       Generic October 1998
$ tpo^?^?^?^?^[[2~^?^[[F
tpo: not found
??: not found
: not found
?: not found
[[2~: not found
?: not found
?: not found
[[F: not found
$ top     
top: not found

At any rate, I just went put four copies of Seti@home on the server (they're really doing a number on the data chunks.. looks like about 10% every 30 minutes). Some Oracle guys will be coming in eventually to actually set up the database, but they want to have a good idea how it will be used before doing anything. Of course, nobody has really figured that out yet...

I have mentioned previously that I was working on a small NeXT box that has been gleefully serving web pages for years. They've been trying to port the backend of this thing over to NT for nearly the same amount of time. From what I can tell, it's just a medium-sized pile of Perl and TCL scripts. I'll have to see if it's hard to port it over to Linux or not. The box is not exactly in my jurisdiction, but only me and my boss are the only people around here who really seem to know how to administrate Unix-like systems...


Hmm.. I think I may pick up that 60 Gig Maxtor drive tonight, or perhaps tomorrow sometime. Of course, what sort of filesystem do you put on that thing? I need journaling if it's going to be that big. I see three options: ReiserFS, Ext3, and GlobalFilesystem. I'm kind of leaning toward GlobalFilesystem, just because I don't have to travel far to physically beat on the developers if something goes wrong ;-)

Diary entry 15


I just had a light bulb burn out in my room. The other one had burned out two days ago. I need to invest in those long-last bulbs, though I bet the ones I get will only last 6 months. (but then I could sue ;-)


The sun was out early this morning, and I almost got up. I really should have gotten up, because the sun disappeared again and I ended up feeling really lethargic.. That must be why I've been having trouble getting up lately -- it's been overcast in the mornings for quite a while now..

I've really got to look into this printing stuff a bit more. I saw reference to Grant Taylor and more CUPS stuff. The Linux/Unix printing system really does need to be revamped. I mean, I don't want to define a new printer in /etc/printcap just to be able to run at a different resolution or to print in color or whatever. This is something that has been missing for a long time. Of course, what's the best way to do this? You don't want every program that wants to print to be dependent upon a particular graphical widget or anything (of course, you don't want it to be dependent on graphics at all).

I don't want more printers to be using proprietary languages, and I think that has been named as one of the detractors to CUPS, whether it is true or not. Of course, perhaps the truth about that rumor is the fact that device manufacturers don't know how to make a `driver' for Ghostscript. AFAIK, all of the drivers are compiled into that program, which makes the distribution of drivers very difficult. Regardless, it sounds as though CUPS can help a lot on PostScript printers (using the PPD printer definition files), especially those that have extensions for landscape, portrait, double-sided printing, stapling, collating, etc. It's a heck of a lot easier to do those fancy printing modes by clicking a button and having your program insert the correct commands rather than going and editing the postscript output by hand..

Hmm.. kind of on the topic of printing, I wonder if there are yet any programs for reading the status of my Epson Stylus Color 640. I'd kind of like to know how much ink is left. I had thought that printer would be very nice, but it's really slow. The heads can get very jammed up with ink, making it impossible to print anything that looks good at all without going through three or four cleaning cycles.. All of these things that I have to buy.. must...resist..temptation...to...spend...money...

I do really need to go out and get a new hard drive. They're selling some pretty nice drives at fairly decent prices just a few blocks from where I live. I still have to dig around the web and see if $210 for a 30GB 7400RPM IBM drive is too much. Or should I just go whole-hog and get a 60GB 5400RPM Maxtor for $280? If that's twice the bit density.. would it be faster than the other one? Hmm...

Late Afternoon

Boy, this is getting to be a long entry.. Oh well. I saw gtaylor saw the previous version of this and gave me some pointers. Thanks for the info..

I've been fighting with CUPS this afternoon. It just doesn't want to behave on my system for some reason. I think the backend filters are getting passed the wrong arguments when they get called. I'll have to investigate that further (and wonder why nobody else noticed..) I suppose the mailing list holds the answer.

Grr.. I'm installing CRSP data on a Linux box here at work. Of course, Linux is not a `supported' platform (not that they seem to support their other platforms very well). They have this ungodly Bourne Shell install script that is really, really crummy. Of course, Bourne Shell doesn't have functions, so you have to cut-n-paste whenever you need to reuse code. Oh fun. And the idiots have to go chmod 555 all of their data. What's up with that? You should use 644, morons! Blah.

Anyway, I have to figure out the best way to lay out the directory structure so that the researchers that use it will actually be able to find the stuff they need...

I've had a highly unproductive day so far. I've been reshuffling documentation for our servers. I think I've gotten rid of all of the old crud. I'm lucky that I didn't throw out one section -- there was a server that I thought had been converted over to NT. We needed the root password.. It turned out that it is just partially converted. It's an old NeXT box (!?!@#$*@%) that runs what is apparently a fairly popular web site. At any rate, the filesystem had filled up from log files. Someone must have tried redoing the main index.html file, so the system truncated the length down to zero when they tried to save it... Oops. Thankfully, Google's cache brought it back to within a few months of it's previous glory.

Still haven't had much time to work on any Free Software. I really need the day to be lengthened to 36 hours...

I hate Solaris. Well, maybe that's a little strong. I'm just so used to Linux, and it bothers me when things aren't where I expect them to be. Oh well, I'll get over it.

Saw that links browser. I'll have to try it out sometime.

Work is pretty slow, though it's not like I don't have anything to do. I really need to reinstall WinNT on one of my boxes. I suppose I may actually have to sacrifice some RAM from one of the Linux boxes I have -- I think the NT box only has 32 right now, which means it's dog slow. Personally, I don't need NT, but I have users who run it on their own systems, and I need to test out software for them (PuTTY and other stuff).

Lately, I've been organizing a lot of documentation. Clearing out old cruft from The Big Manual that we have for all of our systems here. At some point I actually have to use it to re-install a database or two.

I saw David Boies on Charlie Rose last night. He was talking about the Napster case, and was very good. I knew that he exaggerated some things, but he cited the 1989 Audio Home Recording Act when saying that it's OK to sample music and to share music with your friends noncommercially. I think I'll have to do some reading. I haven't decided if I'm going to buy any music this month, but my musical mind seems to be withering. I think I'll have to go get a few (or a lot) of CDs.

The commercial advertising actor strike is really bothering me. Very few ads are being made, so advertisers do not have any variety in their ads anymore. My mind is going numb after seeing the same ads over and over and over. It's becoming sickening, and I've been avoiding TV.

I went and rotated LinuxSecurity.com's Linux security quick reference card. I couldn't get it to print nicely when dumping the PostScript version directly to our printers. A simple `-90 rotate' and some tweaking of a `translate' command and everything was peachy-keen. It doesn't display right in gv, but that's not a big deal.

My job today is to set up all of the Novell print queues on one Linux box. Then, everybody can print with Unix lpr/lpd pointed at the Linux box. Saves me the trouble of setting up IPX on all of the client boxes, but that's just two lines in /etc/rc.d/rc.local anyway..

Well, it's the end of July. August cometh. The Minnesota State Fair will be starting soon. (It's the first or second biggest, IIRC. If it's second, it's probably only behind Texas or something. Not bad for a state of 5 million) That's cool. School will also be starting in about a month. That's not so cool. I also need to exercise and practice for Marching Band.

Stress is building. I wish I had a gf to help me relieve that, but maybe having a gf would just make life even more stressful...

Hey, djcb, fix your last entry -- you're missing a double-quote mark..


Went to the Twin Cities ``Rockin' Ribfest'' today and got a half rack of Roscoe's ribs, some fresh lemonade, and a Dove bar. All for $20.. Eesh. Yet another one of those *fests that that requires you to buy using tickets rather than paying in cash. But it was good.

I rode my bike there and back. There's a fairly nice half gravel/half asphalt road going along a railroad right-of-way. The road basically goes under all of the traffic, which is a much more pleasant way of getting places, IMHO. However, the road is suffering from potholes and washboarding. My arms were getting thoroughly tenderized for a while.

It would be really nice if the University would put a real bike path down there and connect it to the new bike/pedestrian bridge they just installed. One of these days.

I haven't gotten around to working on my bus schedule thing much today, though maybe I'll do something after Futurama and The Simpsons. However, I was pleased that I actually got Evolution to compile and run. Now I've got to try my luck at Nautilus..

One of the things that I worry about with Advogato is that the diary area could be prone to DoS-ing of various kinds. Individuals could presumably plaster diary entries all over the place, or post pages and pages of junk at a time. I imagine there are some checks for this already, but there's almost always a way around it.

Oh, I have a few mail-related problems. One is that I just started using procmail to filter my mail. Locally, I run an IMAP server so it's easier to view mail with disparate clients. Also, I think my mail actually gets loaded faster, as Netscape is not the most efficient at reading/parsing mailbox files. Anyway, I still have to find a way to get procmail to notify the IMAP server that it just dumped mail into a certain folder. Then, the IMAP server must be able to notify my client that there is new mail. This isn't happening automagically, and I'm not sure how to get it to work correctly. I suppose I may have to start using a different IMAP server or something.

In other news, I need to find a good way to consolidate mail at my workplace. Well, my mail doesn't matter, since I'm an admin and can basically do anything I want. However, the head of my department basically wants all mail to go through Lotus Notes (Domino?). I believe it is my task to now find a good way to get Notes to interoperate with my system and the Linux/Unix desktops of others in the organization. I imagine that the server can just run POP or (preferably) IMAP through stunnel. Then, any decent client can read the mail. However, calendaring still remains a bit of an issue, though Lotus is (or at least was) a supporter of the iCalendar protocol, which Evolution is going to support. My only question is, does Lotus actually support iCalendar or not?

Well, I finally got started on my bus schedule project. I'm probably a third of the way through making a decent parser to read in my data (it's full of redundant code, though I hope it will be moderately robust). Then, I actually have to do something with the data and output it into a decent format. I've got a bit of work ahead.

More guests last night. I stayed in my room most of the time, though.. They got in late, and I was busy coding. Bumped into them when I woke up, and they were gone in 30 minutes.

There isn't anything good on TV. Perhaps I should plunk down some change and give in to the cable gods.. Oh well, since there isn't anything on TV right now, maybe I'll actually do something useful. Part of the problem with the consumerized society here in the US is that people forget how to innovate on their own. I realized earlier this year that I had become afraid to be curious about a lot of things. There are so many warnings and locked doors around us these days that we forget what it's like to explore. It doesn't help that every corporation on the planet wants to patent everything they touch these days..

Hmm.. My bus schedule parser sucks.. I think I'll have to rewrite it..


I tend to forget how useful ANSI colors are. They're wonderful for debugging. Especially when you're trying to find a needle in the haystack of pages and pages of output..

Hmm. Need to find a good, clean (and hopefully libre) method of getting Reflection X to work with SSH. Right now, it is heavily dependent upon Telnet or the r* commands.. blech. PuTTY is great for just being an SSH client, but if you need to tunnel X sessions, forget it. Perhaps I should just avoid tunnelling and go for direct X connections (ie. Set $DISPLAY to point to the client system rather than server:10 or whatever)..

Anyway, my boss gave me an Internet Security Scanner report about our servers yesterday. ISS guessed the SNMP community name and could change the system configuration. Oh yay. I went through and disabled quite a bit of stuff. At some point, I need to find all of the potentially sharp objects on the servers and make sure that they are safe. (mostly SUID root executables, but perhaps compilers and assemblers as well..)

We have a Linux Mandrake box that had some pretty nifty security stuff built in. I wish RedHat would do that.. However, Mandrake seemed to go a little too far in some places. I mean, should the /usr mountpoint only be readable by root? *shrug* At least it didn't start every service on the planet when I first booted it up. Of course, that system is the fallback fallback. Well, it will be primarily serving NFS shares. Secondary function is being a fallback NIS server. Tertiary function is being the fallback fallback shell server.

Anyway, the security report wasn't terrible, but not as nice as I had hoped. Of course, my Unix boxen were the only ones that didn't have the problem of predictable TCP sequence numbers (if they are not predictable, it is very hard to do complex IP address spoofing). The Novell servers were the worst, with ISS getting 100% of it's guesses correct about the sequence numbers. Most of the NT boxes were around 60%

I still want to work on my bus schedule proggie, but I don't know when I'll get the time. I can't live without my 8-9 hours of sleep (compared to most techies, I'm a total weenie). I might be able to live with less, but then I'd need to be able to sleep in until late (11AM or so). I'd also love to do some work with weather-related programs. You know, something that would send me a message if there's a Tornado Warning or something. I guess I just need to find a decent data source first.. Also, I hope the (U.S.) government will make NEXRAD radar data available publically when the radar contracts expire later this year. Getting hour-late images from Yahoo and other places just sucks. Besides, wouldn't you like to be able to zoom in on the images just like your forecasters do on TV? Or maybe make your very own 3-D flythrough?

S.5....T   /bin/ps
S.5....T   /bin/login

Heh. That's good.. This box has been broken into before, I guess nobody bothered to clean it up properly..

I've just been going around the different servers and setting up NTP daemons and fiddling with other stuff. The Linux boxes here have been surprisingly stable. One of them has been up 194 days at this point. I almost rebooted it yesterday, but then I realized things were behaving strangely because the automount daemon had died somehow, but the mountpoint hadn't been released. A simple umount did the trick. Maybe that thing will run for another 194 days. I just hope that it will come back up after it finally goes down -- we don't have a display or keyboard on the thing ;-)

Anyway, I'm hoping to work on a new project. I want to make a simple program that will let me input simplified bus schedules and then display them on a webpage (or output to a text file, whatever). It's sort of an effort to keep the campus busdrivers `honest'. The city buses in Minneapolis are very good about being on-time (IMHO). The campus buses for the University, however, vary from their prescribed times by quite a bit. If they could just get down to ±2 minutes (that's a 5-minute window), I'd be happy..

Hopefully, I'll be able to make something simple, yet powerful enough to handle the weirdness of big-city bus schedules..

8 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!