Ten Years of MZ Online!

Practical Advice

Extraordinary Ideas

No-Nonsense Solutions

Saturday, May 17, 2003

May 17, 2003


You read correctly, grasshopper. smile The (new) server box lives...RedHat 9. And it was good! wink

I'm pretty sure I figured out what the problem was with the installation bits -- it had to do with the CD-ROM drive. For some reason, perhaps the ages of the drives (although neither I tried (which ultimately failed) were that old -- a few years -- 2-5 max), the Installation CDs would 'test out' properly (meaning the data itself was good, according to the installation program) but would randomly (in mid-installation) miss a bit or something, causing the whole mess to go boom...in a very bad way (fatal error terminating installation).

Tonight, I first brought down the entire network to try the CD-ROM from the existing server box. That CD couldn't even find the boot image of the bootable CD. Guess that wouldn't work. So I brought down my personal machine as well, stealing the regular CD-ROM (not the CD-RW) and testing it. Strange -- it worked just fine. In hindsight (20/20), I guess a CD-ROM drive cleaning perhaps may help...and I imagine I'll try that for the second installation of RH9...sometime this week hopefully.

While I had three of the five machines lying in pieces at the same time, I took the rare opportunity to vacuum. smile Nothing like a 1.5HP shop-vac to clean out a computer case or three. Compare it to your favorite pen, lost in space...getting sucked into a black hole -- it happened quick...and thorough. smile

Anyway, I've (base) configured the machine (new server) to boot and get on the network. I didn't bother with DHCP stuff, since the DHCP server will move to that machine eventually...so I set network properties manually. BUT, I did get everything running on the 100MBps link. Another big thing.

In the next few days I plan on migrating configuration files to the new server and shutting down services as I get them running on the new box... When that's all done, I'll copy the last of the necessary files (as backup), and wipe/reinstall RedHat on the old server box (as bare minimum) so I can use it explicitly as a router/firewall/timeserver box.

This post was upgraded to the MZ Online Blog on 8/20/07

Friday, May 16, 2003

May 16, 2003

Major Change in Site Look!

I started tinkering with it last night, and anyone who's been here in the last 24 hours or so would've definitely known I was up to something. At different times throughout the day, things would've been very difficult (if not impossible) to read or decipher...for anyone who happened by here during this time, my apologies go out to you.

The major changes you'll notice:

  • Solid, concrete look to all portions of a page
    • I brought together the sidebar with the top navigation bar and also with the lower information bar.
    • Three sides of the page are outlined in the maroon color of this scheme. I felt this was good to 'tie together' all of the (formerly somewhat separate and confusing) bits of the page.
  • Color Scheme
    • I made some minor changes to how links are colored (and how they behave)
    • I also adjusted header/footer text to contrast with the new maroon background

For those who don't understand how the TWiki operates, pages are essentially split into three parts

  1. header (the part with my head shot, telling you where you are),
  2. text (this stuff you read), and
  3. footer (the bit with the copyright information at the bottom)

The software dynamically creates the header and footer for each page, and to some extent (depending upon the user) can dynamically create the text portion as well. When I edit a page here, I only adjust the portion contained in text.

However, for this site, I've got the following individual page bits:

  1. header (customized from stock TWiki look),
  2. sidebar (a separate page I edit to modify links on the left side of the page),
  3. text (regular editable content),
  4. lower navbar (the list of links at the bottom of the page (stuff with the devil image)), and
  5. footer (also slightly customized from stock TWiki look)

Thanks to the invention of CSS (Cascading Style Sheets) and TWikiSkins, I can really customize how an entire web looks. Better yet, once that's been done, I can easily change how the whole thing looks in the event I need a new color scheme or something. Since I've essentially split up the TEXT portion (see above for definition) into three individual pages, I can easily edit the navigation parts separately and have them go into effect site-wide immediately.

As all the changes I've made are on a template/general level, I haven't (and don't plan to in the near future) tested individual pages for compliance. Since they're all generated in the same way, my changes (as long as they worked on one page) should be in effect site-wide now. The only pages which might not be perfect would include the search pages (it uses a different template)...but there again I checked it...and it seems to look like everything else...

With that said, if you notice anything particularly strange or bad, please let me know about it! Thanks, and enjoy!

This post was upgraded to the MZ Online Blog on 8/20/07

Thursday, May 15, 2003

May 15, 2003

No DICT for you!

For some strange reason, I've been unable to access dict.org for the last week or better. DNS queries are returning the correct information, but the trace of the path gets lost after about 15 hops...and can't go any further.

What's more strange about this is that I can access dict.org from mzonline.com (the machine). So there must be some strange link down somewhere... Beth pointed out that the dictionary queries weren't working initially...and I just assumed it was our end...but it's not.

With that said...and with the (eventual) inclusion of a second server box on the network, I've decided that I could just as well download and install the dictionaries themselves (the indexes) and query one of our local machines instead of dict.org, leaving it as a backup if for some reason it became necessary.

The problem: How do I find these pieces without (easy) access to dict.org? The answer: search for a while online...and it will come.

I've found a mirror of dict.org which gave me enough information about how to actually get access to the indexes. I don't need the client/server software (that's already taken care of), but I need the actual files these things read from. It'll be several MB worth of data, but the new server will live on a box with a 20GB drive...so I think I can spare space. And if not, I'll just run it off the older server box which has about 7GB of space.

I really need to invest in some more (large) drives... Ah well. It's not like I'm storing terabytes of information or anything. Hell, with all the drives on this home network, we've got about 75GB of drive space...of which maybe 1/3 is being used currently. The main problem is that this is stored between multiple drives (in the case of the older server, it's 7GB between three physical drives).

Five years!

I just returned an electronic survey for my 5-year high school reunion. Strange. Doesn't seem that long, yet it does. I'm kinda stoked about it, actually. For the first few years, if I ran into a classmate, it wasn't that big of a deal. Now I (still) occasionally run into one of them...and it's kinda neat...if only brief.

Ah well, that's all for now...

This post was upgraded to the MZ Online Blog on 8/20/07

Wednesday, May 14, 2003

May 14, 2003

The WebSearch is over!

Or something like that...

This evening I made the aforementioned (previous news item) changes regarding the search skin for Matt's Home on the Web. It was a relatively simple adjustment, really...just copying the style stuff from one template into another, making appropriate adjustments to reflect the different nature of the script.

Anyway, in addition to that, I also simplified at great length the nature of WebSearch in this web. No longer are all the complex options available, but a very simplified set, which only allows searching through topics in Matt's Home on the Web. This made sense to me, as it seems somewhat unnecessary to allow searches through all public webs (or to put searches into BookView...which is really intense)...as well as allow all the confusing (even to me, sometimes) options for searches. Right now, the way it will stay...is a simple text search (or by topic name). After all, there's not all that much stuff here (mostly news items now) to go through. smile

Sever weirdness is still happening... confused Swapped processors (not motherboards) in the new one, thinking that might be the problem. I'm still led to believe that it's probably something else (like the CD-ROM). For the meantime, I'm going to put that project on the back burner.

I'm still thinking about the software page (mentioned in a previous news post) bit. I've downloaded all of my stuff from epoxy (one of the bigger backend Computer Science Discipline servers at UMM) and am in the process of going through it. I actually dug out the core Java books, too! For the stuff written in Java (namely the RSA stuff, possibly the RatBastard Mailer), I'd like to create either application-like features (e.g. can run without the JDK libraries installed -- self-sustaining things)...or .jar files for download and execution. Upon initial investigation, I'm going to stay away from applets, mostly because all the stuff I've ever written in Java (that's cool, anyway) needs to have access that the Java plug-in (for browsers and applets) won't let you have. But that's another project, and hopefully I'll have most of the more neat stuff I've written published online sometime in the future. After all, I'm all done with school now...so there's no real reason for me to keep it hidden. razz (stick out tongue)

This post was upgraded to the MZ Online Blog on 8/20/07

Tuesday, May 13, 2003

May 13, 2003

The Server Soon Will Live!

The parts I'd ordered arrived today! What does this mean? Eventually I'll have a new server box.

Upon beginning the installation of RedHat 9 tonight, I've been encountering problems namely with the video refresh/resolution. After some time, I got that problem fixed, but then (after going through the package selection process several times) the system first hung up (first attempt), and then (on successive attempts) would return fatal errors when installing random packages.

As I actually did run the media test, this surprises me. I will perhaps have to try a different CD-ROM drive, as it seems like it could be something of that nature. Although, as an alternative, I could install the bare minimum stuff and manually install packages with RPM (which might not be a terrible idea after all).

But I realized shortly after beginning this endeavor that more in-depth planning would be necessary. Namely, which services am I moving to the new server box? Which are staying on the old box? Do I want the new box to be the gateway machine?

My initial plan was to move most every service to the new box and use it as the current server/gateway box, leaving the old server as a backup machine just for storing and making backups.

However, my plan now is to probably do something like this:

  • Install new 10/100 NIC in existing (old) server, configure, and remove one of the old 10bT cards (replaced by the new 100bTX).
  • Set up new server with most services (yet TBD). This would include everything currently on the existing server box, with the exception of the ipchains/iptables stuff and possibly time service.
  • Get services configured and working correctly. This will include complete dictd support (no more querying), DNS and DHCP, amongst others I've left out.
  • Make last backup of configuration files necessary for the existing server.
  • Wipe & reinstall existing server with RedHat 9, installing bare minimum features and only configuring iptables, time, and whatever else I feel needs to remain public.

The entire purpose of having a second server box (aside from the significant gain in speed) is to keep private things private...and minimize the number of 'holes' in to the network. Furthermore, I'm hoping to build a more robust firewall this way.

But, this is a journey just begun...

This post was upgraded to the MZ Online Blog on 8/20/07

Monday, May 12, 2003

May 12, 2003

The Resistance: 2 The Enemy: 0

After some more (final) work on the picture display script (the bandwidth reducer problem-fixer), I've added a (minor) security fix and logging support, as well as general 'niceness' and refactored look to the script itself. This was done late last week (Friday, 9 May).

The security fix had to do with adding 'relative paths' to an image query. I considered the fix pretty minor as Apache won't serve anything outside of its document tree structure, so getting to system files wouldn't work in the first place. However, it would still be possible to get to files 'below' the image directory level specified in the script's configuration by using relative paths (e.g. ../../file.name). This erratum has been fixed, and all preceeding ../ bits are removed prior to execution. There is a possibility of another security issue arising which I hope to verify and possibly fix tomorrow. But here again, this would be minor as the script won't actually give you anything unless the referral is correct, so one would have to be somewhat creative to exploit...and still would only be able to gain access to (potentially) other stuff I have out in the web tree.

I also added logging support in a 'standard' format that I find easy to read (and scan). In my logfile monitoring days, the one thing I've noticed is that it can be difficult to notice irregularities in pattern, so I developed my logfile system for the script to make it easy for me to read. This tells me when people have been trying to gain access to files from bad referrals. I have some other ideas for logfile enhancements which I might also implement...which would potentially be configurable.

Done and graduated!

It's official, I'm done with school. Finally. smile That's all I've to say about that. wink

In other news...

I've got parts on order for my new server box (network cards and mice) and server system (new UPS unit specifically for them). I decided to order a few (cheaper) 10/100 Ethernet cards, just for speed. The network topology will change somewhat, and the current router/server box will eventually become a dedicated backup server. I'm going to upgrade the current 10bT cards with a single 10/100 card (it will only need one in its new location, as it won't be a bridge between networks anymore). This will increase speed on the localnet side, good for backup purposes.

The new server box will have a combination of one of the older 10bT cards (for the external connection via DSL modem, since that only supports 10bT max.) and 10/100 for the localnet, providing extra speed with data transfer...since both of our primary machines have 10/100 cards already). I've had an auto-sensing 10/100 hub for several years now...and it's good to actually forsee it getting used more than it has (on the 100 side that is).

I also ordered two new mice. One will go to Beth, the other will go to the new server box. Beth's old mouse will go to the old server box.

Finally, I ordered a new UPS unit specifically for the servers and network back-end. With the addition of another server box, all of my existing battery/UPS protection wasn't enough. So, I decided to purchase an UPS for them. Both servers and the hub will run off of it. This frees up the existing (smaller) batteries to be used explicitly for workstations and printers. Sounds like a plan to me.

And I also hope to get up a Matt's Software page in the not-so-near future. This will outline things I've developed myself or as a group... But more about that later... smile

This post was upgraded to the MZ Online Blog on 8/20/07

My Blog Profile

My Photo
Minnesota, United States
View my complete profile

Miscellaneous Bits

  • One Gorgeous Daughter (Kirstin)
  • An Incredible Wife (Beth)
  • Two Cats (George/Felix)
  • Two Dogs (Koshka/Kurva)
  • My birthday is at the 48851th place in Pi
  • Frank Oz and I share birthdays (Go Fozzie Bear)!
  • Mike Myers and I share birthdays (Party On, Wayne)!

MZ Online Blog: The Archive


Save to del.icio.us Save to del.icio.us W3C XHTML Valid W3C CSS Valid W3C Tableless! W3C WAI Double-A WCAG 1.0
Copyright 1997-2008 Matt Zaske. All Rights Reserved. | Privacy Policy | Terms of Use | Search | Site Map