[Most Recent Entries]
Below are the 15 most recent journal entries recorded in
|Sunday, December 11th, 2005|
I couldn't resist and deployed wordpress
and started posting to dammit.lt
. So long, Livejournal! It was nice here, but I feel better at home :)
|Tuesday, September 20th, 2005|
|a quiz for sysadmins
What happens, when 2 out of 3 major database servers fail for a world top-40 website?
Wikipedia during peak loads (oh, Tuesday), was working on only single db server, that served 6000-7000 SQL queries per second. For curious ones, this server is dual-opteron 16gb ram box with Linux Fedora and MySQL 4.0, all tables use InnoDB engine.
Now we're kind of trying to wake up both other major servers (ariel died because of some RAID issue, adler - bad RAM, bad luck, on-site support will help us soon). More database servers are ordered and might be deployed in two or three weeks, so in future we will have somewhat better way to deal with such overloads.
|Saturday, September 17th, 2005|
|Sunday, September 4th, 2005|
|five years in single company
... today is 5 year anniversary of my employment in current company... there were several name changes though:
- DELFI Internet
- MicroLink Data
- MicroLink Lietuva
Oh wait, telecoms announced quite ago the possible merge/acquisition ;-) Five years did change a lot, what were the changes in YOUR last five years?
|Thursday, September 1st, 2005|
|cheap labor from eastern europe
I just finished my voyage from aKademy
. My flight Malaga->Madrid->Prague->Vilnius was rerouted to Malaga->Madrid->Dublin->Vilnius. The Irishman who sat in same row started a conversation in quite a funny way: "How was your working in Ireland?". See, 90% of that plane were Lithuanian immigrants seeking for a better life...
And, haha, when after passing about twenty security checks in one day they asked me to take off my shoes (on the last one!), I thought it was a joke. It wasn't. Take off your shoes, sir!
|Thursday, August 18th, 2005|
|Wednesday, August 10th, 2005|
|Wikimania: The community miracle
It has been several days after Wikimania, the wiki conference in Frankfurt. It was awesome. It was flawless. And as it was done for the first time - it was absolutely sincere. Ah, right, it had one major problem - it established really magnificent standards for a wiki conference. For any conference even. And we'll have to try to achieve that next year again.
I had two parts of it. The first part was small and was developer meetup. It was absolutely amazing to see how fast ideas can explode from a simple word in the air to an implementation of a concept people were just dreaming about. It was sure great to meet all people in a small crowd, meet new ones, build the trust and fun :-)
The second part was big one. Lots of people, lots of press, and sure, lots of community feeling. In those days you could hear all kinds of neat utopias, saving us from future dystopias. You could be drawn into current and future revolutions or just evolutions. You could find out joys and wishes of other side of our planet, that usually is somewhere just out there, but behind a firewall of culture gaps. In those nights you could just relax and be a part of most wonderful community out there. Ah sure, I slept for 14hrs when I got home. Sleeping in Wikimania is for sissies. There are much more nice ways to spend your time there.
Anyway, thanks to everyone who came. Thanks to everyone who made it happen. Will see you again next year. And sure, for those who were not there - watch some photos on Flickr
. Current Mood: enlightened
|Thursday, July 14th, 2005|
For a while we'll have public access points in main streets of Vilnius' oldtown. This means that this summer you can sit in quite a lot of squares and parks, nice caffes or just on a pavement and have quite speedy net access. No more wardriving in city center :)
Service is provided by Lietuvos Telekomas
, a company that my current employer is merging to. Well, let's check how fast I can go 'home':
traceroute to medium.cs.microlink.lt (126.96.36.199), 64 hops max, 40 byte packets
1 192.168.1.1 (192.168.1.1) 5.663 ms 1.622 ms 3.846 ms
2 213-190-60-9.telecom.lt (188.8.131.52) 12.361 ms 15.624 ms 11.901 ms
3 212-59-31-153.ip.takas.lt (184.108.40.206) 10.679 ms 10.928 ms 10.236 ms
4 212-59-21-166.telecom.lt (220.127.116.11) 11.166 ms 11.966 ms 10.771 ms
5 f0-0.cr0.vln0.microlink.lt (18.104.22.168) 29.725 ms 18.380 ms 35.076 ms
6 medium.cs.microlink.lt (22.214.171.124) 12.553 ms 16.832 ms 15.786 ms
Not that bad for public/free access in Eastern Europe. And sure, b/w:
466398 bytes received in 00:23 (231.72 KB/s)
Well, my office is now in street :)
|Friday, July 8th, 2005|
|In information age speed of light means SLOW
There are two kinds of people who complain about speed of light. First ones are astronomers and others are distributed application developers and both groups are sad because they have to wait for information.
There is a saying "never underestimate the bandwidth of station wagon filled with tapes". Exactly same issue is with fiber optics. You can transfer tremendous amounts of data in seconds, but it will still travel at (or slower than) light speed. It will take your data 100ms to cross the Atlantic ocean and return back and you can already blink at that time. You may request another packet, and you already have to blink yet again. Oh right, you need handshake at first - add more blinks. It might be easier to ship your data in a marine container.
It is not the largest distance on our Globe. Data has to reach Australia, Asia, Atacama and Antarctica, 200ms, 300ms per single round trip. It simply means that if you request data across the ocean, you have to do it either once or never. This is why distributed/global application developers hate speed of light.
Even 100km distance is large enough for synchronous database mirroring - every transaction will take at least several milliseconds and applications will have to be designed with lag in mind.
Or... we should escape from those slow constraints. We need real information age :)
|Wednesday, June 8th, 2005|
|Wikipedia enters Seimas, Lithuanian Parliament
Well, as I talk to everyone about Wikipedia, I did evangelize Public Relation folks in our (MicroLink
) office as well. They were really cool and helped to arrange presentation in Information Society Committee at Seimas
Parliament of Lithuania.
Though, we did discuss this a bit months before in Lithuanian Wikipedia community as well as international folks I did feel quite firm and prepared a message:
- Wikimedia Foundation does not ask for support of operation.
- Wikimedia Foundation _supports_ operation of all content platform.
- What we need is hands-holding and various education initiatives in Lithuania, and we want to help there.
- What we also want is some clarification and legislative assistance with copyrights and government information.
- And sure, we want to evangelize all people to build information
society and information itself.
The day before already sparkled some threads and discussions. My company (MicroLink) was actually considered to be Microsoft, therefore it was told in newsgroups, that Microsoft is now pro-wiki, pro-open source and pro-community.
Also people were thinking that we might be asking for money. Well, we always leave possibility to donate. But we never allow to attach any strings.
Anyway, there were various people in there. There were other MPs in my presentation, and some of them did catch the idea pretty well and explained it to others. Some were still arguing that if anyone can change stuff, then you can't trust the information. I could be seen not that sane there :)
Anyway, it was worth to go there, as some people already showed enthusiasm and initiative, and it won't be the end of our actions. There was quite a lot of enthusiasm from reporters as well - these guys sure do use Wikipedia and some even edited it. One has been involved in revert warring, so our conversation was quite tense :)
Right now I'll be trying to overlook media coverage. It might be nice. It might not. We'll see. Current Mood: ecstatic
|Wednesday, April 20th, 2005|
|habemus effect aka Popedotting
As Slashdot effect
does not have any effect on Wikipedia, new ones appear. Like the new one - 'Habemus effect':
White smoke! White smoke! Hopefully not too much of smoke from servers :)
Ah, and Reuters report on news.com
regarding it. Long life for Benedictus XVI, we've got to upgrade hardware! :)
BTW, some squatters have already registered domains of possible pope names and are selling them. There's one, who did hit the proper one
. Current Mood: amused
|Friday, April 15th, 2005|
|lies, damn lies and statistics
Questions like 'how much hits do you serve? how many servers are there? what bandwidth? how many lines in software?'
et cetera are quite common for those, who are interested in various IT projects. Usually getting a number brings satisfaction and you can knowledgeably nod your head and say "Cool!". What do people often miss, is that you can't measure everything in plain numbers.Without cost estimations, counters are useless.
Single pentium server may serve thousands of single-pixel images or text documents, fill wire speed (gigabits!) with some bulk data. But in project you shouldn't care about theory - you need information about your tasks. This is often brought by 'quality, not quantity' slogans.Single slow task may destroy whole picture.
You may have really lots of efficient code, but single inefficiency may bring your whole system down. Imagine an elephant walking in a highway. Should you count elephants or other entities in there? Is it possible to measure both?Real value is productivity.
In case anyone wants to benchmark any project, they'd start with productivity, brought by it. All performance details have to be considered only after whole picture is known. How would you account productivity?
And yes. I'm obsessed with numbers myself. But that isn't lethal, is it?
|Tuesday, March 1st, 2005|
|I'm no longer youngest, at least here :)
Well, I started my IT career, when I was 14 year old. I did on-site troubleshooting for troubled internet and UUCP users. My credibility was often doubted, I was too young. When I was 15, I started working as system administrator for an edu institution with ~100 staff. Well, I was youngest there, and average age was ~45-50. Later (at 16) I moved to an IT solutions company, and there I was youngest as well (what a surprise! :). Other guys were at least ~25 year old then. We did lots of interesting work together, and I did various stuff at large factories and banks. Did they trust me? :) Then, being 17 I moved to my current job. It happened at September, 2000. As my company was international one, I was the youngest in 500+ staff. Here I did lots of system work, and in one CNET interview
I even told my age, I thought it was fun. It was always fun. About two years ago I lost my 'the youngest' title in international structure, but retained it country-wise. So... today, being 22, I did find out, via phone call and CEO's internal memo later, that there's younger one than me. The words in the memo were "I'm very sorry, Domas, you can't always be the youngest". I'm getting old.
Anyway, no longer miracles of the youngest. Time to start working. Cheers. :) Current Mood: crazy
|Wednesday, January 19th, 2005|
|welcome to the world of nutrition!!!
Today I've entered new era in my life. I started using my new kitchen in apartment I've been living over two years already. I could survive, as I've got a superb Chinese restaurant at the ground floor. I could survive, as my neighbors lent me a microwave oven. I could survive as well, as I could always balance between fast food home delivery and dozens of caffees and other places, where you get hot or cold food.
But anyway, there was empty corner in my living room, that has been labeled on maps as a 'kitchen'. And all I had there before was just a pile of various ingredients, used for some simple cookery. I did think I'll have my range installed on friday and I even bought a frying pan then (that took over half an hour at ~20.00 while I hadn't eat a bit whole day). Damn! Everything was not connected yet. And today I've got everything installed. Time for master cookery!
I had some frozen fish fingers (that stuff is made out of fish tails, heads and eyes, maybe something else). I had some olive oil (well, I thought olive oil is good one ;-) I had some hot ceramic glass surface. And I didn't cook anything for five years (since I've left my parents' home).
Well, at first I tried to understand, if that ThermoSpot technology does work (the red dot should glow in case of ReadyToRoll condition). I did wait for a while and didn't notice any difference. I even did put cover on. Afterwards I did put some oil into pan. After a while it didn't start to boil (should it?:) Then I did put fish fingers. There was some nice sound of boiling oil and stuff. Yay! At least I've got some progress. I did set timer to 5 minutes. Afterwards I decided to turn those fishpieces over. Those were quite brown, and I quite liked the smell ;-) Then I did wait for one minute, turned everything off. Those fingers were put then into dish, and... covered immediately with thick layer of Bolognese sauce. Voila! At least it looked (or was hidden) nice ;-) And it ended up in total customer satisfaction. My first try to enter the world of proper [sic] nutrition didn't end up in thrash-bin.
So now I wonder, what will be my second dish... Help please ;-) Current Mood: amused
|Monday, January 17th, 2005|
|blogger age distribution
Mhm, it's a strange situation - I'm entering blogging not in my teens. According to LJ age distribution, I'm already on the older part of the trend. I've never done anything this way. I'm getting old. And I'm far behind current tech trends? :) It's amusing. Current Mood: tired