Everytime I walk into a library I usually torment a member of staff (I don’t think librarians in the classic sense exist anymore) by asking them about Wifi. I’ll argue it should be free, like the rest of their services.
One member of staff from the State Library concluded an argument by claiming the library isn’t an ISP!
I wrote a simple Web application to tail log files.
For the meantime the files on those accounts will be still there, though my mail won’t be redirected. So contact me via my iki.fi account. iki.fi supply a great service. Lifetime redirection, so I doubt my email will ever change.
Anyway, my shell accounts in Finland were about 1000 times better than they were in my previous University Bath. I remember in Bath getting a 20megabyte quota on a stupid Sun machine. OMFG…
Initially in Helsinki my quota was a soft 2G limit. By the end when I worked for the department it got quite rediculous.
It wasn’t just the space, the Internet speeds in Finland are stupidly fast. As I left the department they had Cat-6ed Gigabit ethernet and it could well be called Gigabit Internet. There wasn’t any pay-as-you-go limits with Finland either. I remember in Bath the IT staff begging us not to cause traffic. In Australia though, we’re in the dark ages with people having to pay for inbound data. Argh!!
The CS environment wasn’t Debian. It was a RedHat hybrid called CSL. Jani Jaakola was behind this and I must say he is easily the best Systems Administrator I’ve encountered. I had a stable desktop, that was far more bleeding edge than Debian stable at the time. Jani by himself did what Helsinki IT and Helsinki University of Technology(HUT) departments couldn’t. A decent linux desktop.
Another honourable mention is Pekka Niklander. Pekka is the best at what he does. I would regularly visit his office for help as well as a chat. He would be the one who answered my numerous emails to support. I have a lot to thank him for. It is such a pleasure not to worry too much about hardware and the environment and get down to research. Thanks again Pekka.
Damn, I am getting all emotional.
Today I am using Californian based Dreamhost for about 80% of my hosting requirements. I was recommended to them by Hixie and I thank him for it. I rely on Jamie’s London based DSL-hosted AmPC running unstable for some things still. At Dreamhost they’ve recently upgraded to Debian sarge, which is nowadays pretty current. They also do hourly, weekly and monthly backups which is super sweet and is on par with Helsinki’s setup. My online digital life comes to about 7G of 16G that my Dreamhost plan offers. 7G basically includes all my projects, mail and pictures. So I am fine for space and I never even get close to my 210G monthly bandwidth limit.
At work I am keen to move away from white boxes to something say listed on Red Hat’s HCL.
* Unable to find a midi tower with my specs in mind
* Unable to quote over the Internet
* Take forever to quote from what I’ve heard
* Stupid promotion with renting a server…
* Too expensive
* Unable to quote over the Internet
* Stupid size of company questions. I KNOW THE SPEC THAT I WANT.
* Minimum requirements of software. FFS…
* Too many unneccessary questions
* Awful Website, awful URLs, awful navigation…
* Microsoft 2003 server
* Terrible Website, terribly long forms, URLs make little sense
* Asking me for too much information like the size of company and applications! FS.
* Getting put on hold and transferred to two different sales people and then onto an engineer in Bangalore
* Still takes seemingly forever to quote
* Seem to settle on a SC1420
* Eventually get an email quote with PDF attachment with 200AUD knocked off the quote given over the phone. I hate these annoying Dell promotion codes.
* Difference in RAID-1 solution and non-RAID solution is like 1000AUD. That’s insane.
* Trying to get me to buy 4×512 RAM instead of 2×1G. Honestly…
* Seem to be unable to refer to a Web form to complete payment. FS. Looks like I have to call them up voice if I want buy this. ARGH!
Update: Another email from Dell with a large PDF and survey, saying if we order today, we get 100AUD off! Instead I’ve decided to buy local a rack and cabinet from our dedicated server hosts, Hostcentral. The price in the end wasn’t far from Dell’s quote and it was a more “professional” rack format.
debian$ dict qualms bash: dict: command not found
This happens to me far too often. I can’t be bothered to install dictd and the database files. It is usually quite a mess and I’m not too happy with the results. Instead I use the Web:
I’ve stopped using dictionary.com although it does seem to be first site I think of because of its great domain name.
Now I just need to find a decent thesaurus not from *.reference.com
- Restarting the the Application server is PAIN. You need to do this every time you deploy. It adds at least 30 minutes to a software iteration.
- Hot deployment of Enterprise ARchive*s simply *doesn’t work. Save your sanity and don’t even test it.
- Multiple instances of JBoss in the same machine is INSANELY difficult (allocate 13 ports for each application server x 6)
- How awful are those XML configuration files? My ports-bindings.xml is 1000s of lines long
- The logs suck. From their init.d, there isn’t a date stamp. You’ll have to check the actual logs JBoss generates if you get JBoss running.
- If you have Swing applications that use JNDI. Good luck with the firewall administration.
- Community communication happens on a PHP forum. Professional? Please.
- The Web console. Shockingly bad.
- You’re crazy to run it at as a HTTP server. You’ll be requiring mod_jk.
- Read the JBoss blogs to witness crack abuse. Quote: I love XML
- It’s a pig. JSP recompiles on a restart and all sorts of mysterious slowness for whatever stupid reason.
Remember Debian’s little issue delivering a Xfree86 security update?
- FS, more stupid “Open source licenses” from Zimbra. I can’t be bothered to figure out what’s different from the GPL…
- Uses proprietary Java with Tomcat. Whoa. I’m surprised. I thought this is “Web 2.0” with REST and well… no ugly JAVA
- How does a project keep track of all those .JARs? And keep them updated.
- Build is nice and clean (less than a minute to compile the ZimbraServer’s 617 source files). Although reading an Ant build.xml is a nightmare.
- In the ZimbraServer/src directory there is Perl and a tiny bit of PHP. I’ve never seen that before. I’m very interested in migrating some functionality from a product at work to something like PHP. I even see C source code: ./ZimbraNative/src/native/IO.c
- find ./Ajax/WebRoot/js/ -name .js | xargs cat | wc -l is errr *28146 ! Subtract about 2500 lines of the license header blah…
- Quite a bit of Thirdparty stuff in the release: amavisd, clamav, cyrus-sasl, jakarta-tomcat, java, mrtg, mysql, openldap, Perl, PostFix, rrdtool, sleepycat, snmp. That makes 132M of the 150M package! Should depend on packages, Debian style. Oh well, the IT has a lot to learn still about software package administration.
- Developed in Eclipse, judging by the unwieldy .classpath files
- MailFiltering.doc ?? Urgh! No Word thanks. *.txt is nice though. Keep that up.
- They’re using SOAP. Sigh.
- Anyway for a Java project the code looks very good. I’m learning from this.
This time last year I was in Pakistan. I’m very saddened to see the news this morning about the Earthquake in the region.
I hope the good people I met there are safe. These people have enough hardship and for this to happen to them makes one despair.
Umm, why has the .ssh/known_hosts format seemed to have changed on my unstable machine?
Can’t see the hostname or the IP address.
vi /usr/share/doc/openssh-client/changelog.Debian.gz +/HashKnownHostswrites Adeodato Simó.
I like Bloglines except:
* sometimes how text is resized in the frame and I have to scroll horizontally
* it doesn’t update the feeds as fast as I would like them too in some odd cases
* Russell gave them a bashing
Though after using Google’s “Reader”:
* The UI sucks
* It’s slow
* I don’t get that broad overview that Bloglines seems to provide in their left frame
* It seems to have all the RSS feeds in one column. I don’t read some feeds very often and I prefer it they we’re seperated like Bloglines instead of cluttering up everything.
* Ok, I DO like their keyboard shortcuts.
Recently I invited a friend to Gmail and she got a googlemail.com address which is too long. Argh. Google, come on.
I’m using Gmail for spam filtering, and Groups pretty often. It’s nice when they have a new service I don’t have to create another account and I can just use my Gmail account. That’s a major advantage alone in this Internet game.
Conclusion: I am using Bloglines for RSS reading for now.
And after working in commercial environments with Informix and Oracle, you really start to appreciate MySQL’s quality. No strange quirks or odd behaviour like my Oracle experiences. It works like you expect it to and with Debian packages it is an absolute breeze to maintain/update/upgrade.
The legendary feature that MySQL missed, the feature that seperated the boys from the men a.k.a. Oracle, was Transaction support. It was only recently I discovered that MySQL has in fact transaction support with InnoDB and big players like Wikipedia and Slashdot uses it.
So what could this mean? I’m taking quite a pessimistic approach here as why should Oracle fund development of a product that basically competes with the Oracle Database product? That would be like Microsoft competing as a Web application platform and undermining its core Win32 platform. DOH!
# Infeasibly they could get each developer to change the licensing to something non-Opensource. I think you are allowed to “revoke” the Opensource license (or any license), though I don’t think it’s practical.
# The future Opensource licensed versions of InnoDB are in doubt
# InnoDB goes “professionally” unsupported
# The developers get bought out to not do any work on InnoDB and instead work on something else or go on a long holiday. It only takes a buy out of a couple of core programmers it some projects to absolutely ruin a product. Remember Borland?
# An opensource company/project is founded which works from InnoDB opensource base. I haven’t looked at the source code, but this is always hard to do. Some people from the community would have to super motivated to do a good job here, though the potential rewards are high. The DBMS market is worth a lot.
# People use Postgres instead. Some people on WebmasterWorld have justified this move by citing MySQL licensing issues and some other concerns with drivers. I’ve used Postgres and I’m not going there again.
I think option 5 will be realized despite the difficulties in finding the right people, because the potential gains are so high. Even if just to be bought out again.
Conclusion: Just a bump in the road towards opensource domination.
Notice how the loose ribbon of window packaging keeps the glass clean without water?
I’ve moved my Subversion repositry over to Dreamhost. As that is a shared host I needed to setup Trac via htaccess.
Now I need to get rid of the stupid index.cgi crufting my URL!
Should be same URI structure as my previous host:
I guess I need to do this with RewriteRules from htaccess, but I have never had much luck with them. They’re voodoo. Black magic. I need help.
There should be some resource or group dedicated to solving RewriteRule problems like mine. To de-cruft URLs and maintain a pleasant URL structure.
While using my shell over at Dreamhost I did an instinctive apt-get search and noticed some strange packages. After checking with Debian I quickly realized these must be packages Dreamhost came up with themselves to manage their hosts.
These “re-usable” bags in Australia are a scandal. The plastic bottom always breaks. How is that supposed to be durable?
The less idiot “Bonds” running around the better.
Reminds me of the effective Stasi when they ask you to compile reports on your neighbours and friends.
Spelling on my blog will improve again.
I took a look at Google sitemaps. It doesn’t seem to support subdomains. I use subdomains everywhere.
I thought since I ran my own blog host, that I would have more comment spams than most. My sister tells me she has to spend time deleting “junk mail” from her Google-run blog too. I am getting about 50 spam comments a day and I’ve blacklisted all the obvious stuff. Argh!
Just out of interest I seem to have about 900–1000 SPAM emails in my SPAM folder (which automagically purge 30 day old mails) on Gmail at all times.
As I don't have a credit card, nor want one I investigated paying "Ludicrop Research and Development aka Flickr":http://www.flickr.com/account_order_faq.gne#alternative 24.95USD via International Money Order.
The Western Union money order from Australian post costs 8.65AUD. And I would have to pay additional postal charges including an envelope. Argh, pain. There is at least an additional 30% charge to pay Flickr via this method. Not to mention it’s snail mail slow.
Welcome ladies and gentlemen to the shitlist.
Blogs in this category will contain a little rant about something or someone that’s polluted my inner peace and pissed me off.
Today’s it’s STUPID D-link Australia.
At work we have a faulty modem/router DSL-504G crap that continually needs to be power cycled to get working.
D-link says we can return it on a RMA and get it replaced within a couple of days.
BUT HOW CAN A BUSINESS DO BUSINESS WITHOUT INTERNET FOR TWO DAYS?
What you should have done D-link is this: Replaced the stupid piece of hardware with something that works.
Went through a tricky upgrade from Apache 1.3 to Apache2 on debian-apache.
Somehow the init.d script went missing and all hell broke loose.
Thanks to Adam Conrad for helping me out.
Today I was asked to make a report from an Oracle database.
I thought this would be a good time to test PHP with Oracle. I’m sure there are some Oracle reporting tools and I’m positive they suck.
Hmmm php.net reports that oracle functions are deprecated… I’ll Google. This Oracle Faq tells me I need Oracle Call-Interface functions.
So I snag instantclient-basic-linux32–10.2.0.1–20050713.zip weighing in at just 33 megs. That’s a slim package from Oracle. I wonder if that version works only with a specific version of 10g or any? Oh well, Oracle are not giving me any clues.
After navigating about umpteen pages I then have to log in. I try my metalink account and guess what? It doesn’t work. Off to Bugmenot for some sanity. Oracle’s disclaimer is awful to attempt to read. I just check the boxes and download…
Sadly to my WinXP workstation as there isn’t a unique download URL for me to plug into wget in my putty session. Oracle you idiots. Though, I recall Sun being just as dumb. I now boot up WinSCP and transfer it to my Debian machine.
So I unzip the beast:
Grrr, is this really for PHP? Where is a nice little TEXT file to read?
Google. Google some more. Time passes. Check email. Google. Holy Shit. Do I have to recompile Apache?
Nah, can’t be. Can’t be… Maybe there is a php_oci8.so I could reference in like MySQL in the php.ini. There does seem to be a .dll for windows. Grrr…. Nothing. Maybe PHP5? No, these guys say stick with php4 with oracle.
Oh here’s something. I love php.net’s comments.
Maybe just a PHP recompile. Argh… After reading the comments, Oracle seems such a nightmare to install. I wonder what glibc these instant clients are built against. I wonder if the instant client package is enough. I am going to cry now and come back to this.
Why didn’t I heed my own advice?
Honestly, I tried.
So that makes Subversion close to useless in large computing environments which generally always run NFS to provide mounts.
It’s a miracle I am still using Subversion knowing what I’ve been through. Omg the Berkley stuff still gives me nightmares…
So I moved back my repository to Jamie’s ADSL hosted unstable box.
At least now you can see what I’ve been up to.
A Microsoft machine seems to be infected with some GAIN spyware. I thought I’ll just go and download Microsoft’s Spyware instead.
Except it wants you to validate your Windows with an EXECUTABLE. How scary is that?
I swear I could choose to avoid the validation check in the past. So how do I do it now?
.htaccess: RewriteLog not allowed here
So what does one do?
While preparing alpha Debian packages for WP 1.6 I looked at the htaccess Wordpress generates. It’s just:
RewriteRule . /index.php
I didn’t know you could do this. This makes it so much easier to de-crufy URLs. I started an interesting thread on Wordpress hackers which details this sane URL rewriting technique.