Posts Tagged ‘technology’

Mozy

Monday, October 20th, 2008

I’ve been trying out Mozy for backup for a while (the free version limited to 2GB of storage).

I recently upgraded to the mozyhome edition, it’s only 5$ each month for unlimited storage, but only for ONE computer.. And since I’m cheap and there is no Linux client, I just mirror the things from the server I have to back up in a directory on my windows PC (SVN repository, postgres database, web server config/logs/etc).

Uploading the initial data (22 GB) took around 3 days, but the incremental backups are much quicker.

I like that Mozy stores several versions of the file, so I can roll back to a specific date. They only save the different snapshots of files for a month, but for the important stuff I have SVN which lets me do unlimited rollbacks anyway.

update April 27th 2009
I recently managed to corrupt a 30 MB file and had to restore it. unlike the upload of the data, restoring it (downloading) was extremely fast, got around 6 MB/s download speed :)

update April June 6th 2011
Mozy stopped with the unlimited plans, making me get a 125gb plan at twice of what I paid before. So obviously I’m switching to CrashPlan when my mozy account expires. The Crashplan client is much better (allows it to also back up to a local drive, or at a friends computer), and they offer unlimited upload at the same price as mozy 50 GB.. They also have a family option that includes up to 10 computers. It also seems like you can install mozy on both your stationary computer and laptop, and have the laptop backup stuff to your stationary computer, which then backs up to the crashplan servers. Also Crashplan supports linux :)

VisualVM – nifty java tool

Wednesday, July 9th, 2008

I downloaded a new Java tool today, called VisualVM which was released (version 1.0) yesterday and is going to ship with the SUN JDK starting with JDK 6 release 7.

This is a tool that allows you to view information about the java programs currently running on your computer (or a remote computer). It does not seem to provide as much information as JConsole or JManage, but it still gives the most essential information. Besides the tool is still pretty new (version 1.0), so I’m sure we will see lots of improvement in the near future.

Monitoring provides basic information about uptime, memory usage, threads and some other stuff:
Visual VM Monitoring

The application also has a rudimentary profiler which allows you to see which methods are using cpu cycles, and which objects are hogging all the memory,
VisualVM profiler

The program can also do a heap dump, allowing you to see the size and number of instances of each class. You can also inspect each and every object to view it’s variables and references:
VisualVM heap dump

The profiling part was pretty simple and would benefit from more work, but overall I’m pretty sure this tool(even though a bit low on features/information in it’s current state)  will be a big hit, and a good entry to profiling and application management to many people.

A very nice feature I did not take a screenshot of is a visual timeline view of all the threads and which state they are in, very smooth :)   The software is also suppose to let you view core dumps, but did not get a chance to test this feature tonight.

I hate computers

Friday, June 13th, 2008

Goddamn screen, or graphics card.. or both..

I had to connect my Samsung Syncmaster 225bw to the server so that I could change the ssh port back to normal (had changed ssh port, but did not change the firewall, so I was locked out from connecting to it).

The monitor has been connected to my desktop computer using DVI, but I had to connect it to the server using VGA. However when connecting the screen back to the desktop the screen would NOT work using DVI, it only worked in safemode.. I tried lots of crap, searched lots of forums but nothing there worked. In the end I had to boot in VGA mode (640×480) and remove and reinstall the nvidia drivers..

Now everything seems to work fine again, except I lost 1-2 hours of my life (and maybe a week indirectly by the stress, anguish and hatred towards monitors the event caused me).

SERENITY NOW!!!!

Bye bye tomcat

Tuesday, April 8th, 2008

I’ve been using tomcat for quite a couple of years now..

But when making working on a registration page for a “ferret show“, I got so goddamn fed up with two things:

1) Each time I recompiled some of the background Java stuff (beans + database manager) I had to restart the server after recompiling

2) OFTEN after recompiling, for no frakking reason whatsoever. Tomcat would bitch about being unable to find the postgresql driver, even though it was in the common/lib as well as the webapps lib directory..

The second issue would sometimes take 5-10 minutes of removing the file, moving it back, restarting the server, removing work/tmp dirs, before it suddenly and miraculously worked again.. So I finally had it, and downloaded Resin and had it up and running in a few minutes.. Copied over the webapps, and now everything works nicely, and I never have to restart the webserver.

I’ve had some experience with Resin from my previous workplace, using the java php implementation and the JMS system. And Resin seems to be pretty nice, and switching from Tomcat to Resin only took a few minutes.. Did not tweak the config much and set up SSH, that will have to wait for the weekend :)

Bye bye CVS.. Good morning SVN

Tuesday, March 25th, 2008

After years of faithful service, I made the change from CVS to Subversion (SVN). I installed SVN on the server, and set up the SSH keys on my desktop and got up TortoiseSVN and SVN for eclipse. I used cvs2svn to copy over everything from CVS to SVN (which required about 1-2 hours of removing the Norwegian characters Æ Ø Å from filenames from misc. schoolwork). I like TortoiseSVN much more than TortoiseCVS, but I’ve had some problems with checking out stuff into eclipse. Seems like I have to check in everything as eclipse projects for it to work straight away. I notice that SVN seems a bit slower than CVS (when using it with SSH and pre-authenticated keys), if it becomes a problem then I’ll just install SVN on my windows desktop and mirror the repository on my server.

using a large JLabel in a JScrollPane

Thursday, August 23rd, 2007

I was updating the SendSMS program I wrote, and instead of a JTable I wanted the list of messages sent to display in a better looking and more intuitive way. I decided to just use a JLabel and use html for displaying the sent messages:

JLabel using html in JScrollPane

I had the JLabel in a JPanel with GridLayout(1,1), and the Panel inside the JScrollPane. The JScrollPane always have vertical but never horizontal scrollbars.
The following problem arose: When the message got to long, it would not wrap, and end of message would disappear to the right.
The solution is simple: set the Width of the JPanel with the label to be roughly the size of the application window..

That did the trick, now however the scrollbar never updates when the Label gets very tall.

Apparently you cant just set just width for the Panel, you also have to set height.. again the solution is simple, set the height of the panel to be the height of the JLabel.

However since the JLabel is inside the JScrollPane (at least, I think that is the cause) the getSize and getPrefferedSize will return only the visible height of the JLabel. It took some playing around with but finally got it working by using the getMaximumSize.. NOTE: dont use the width from getMaximumSize() cause it is ridiculously large :)

the code:

historyPanel.setPreferredSize(new Dimension(historyScrollPane.getViewport().getSize().width, historyLabel.getMaximumSize().height));

Now I just have to find out if that way of getting the components height works the same across different platforms :)

google analytics

Sunday, July 29th, 2007

Seeing that I never do anything with my tomcat log files, I tried out Google analytics on both this site and aakretech.com/net/org.

Adding the tracking code in WordPress was very simple, but I had to rename all the files on the Aakretech site to .jsp, and include the footer on every page. I also added

I was pleasantly surprised by the analytics web interface, very nice looking, user friendly and has a wide range of statistics. the highlights where the nice map (supports zooming down to city level) and of course seeing both Firefox and Safari whoop IE’s sorry ass :)

google analytics mapEven a visitor from Kanagawa Japan, Totemo kansha shimasu :)

browersIE(orange) got nothing on the other browsers :D

Current events

Sunday, March 25th, 2007

Haven’t posted much lately.. Things have been pretty busy with work and the moving (only 6 days left, yaaaaay!), I’ve had almost no time for UFO Afterlight or Silent Hunter 4 the last 2 weeks, and I don’t expect I will have much time for C&C 3 when it is released next week.

I took out my scooter today, drove around a bit.. took like half an hour to start the bastard after it being in storage for 5 months this winter. I’m looking forward to not taking the bus to work tomorrow :)

The other thing that has kept me busy lately is that I have been making a new webpage (that combines aakretech.no, .com, .net and .org into a single page). Not quite done yet but most of the pages are in place. Removing the cocoon webapp for aakretech.org from my Tomcat server reduced the server’s startup time from 15 seconds to 0.7 seconds.

Not much more to say really.. Nemi got her new cage, but we wont get it for her until we move next Sunday/Monday
ferret nation

all work..

Sunday, February 4th, 2007

Today I did the most tedious work ever.
Having committed everything on my CVS server using a Linux client for the last years, I accidentally forgot to mark stuff as binary. As a result every binary file (like pictures, datafiles, etc..) get fucked up if I check them out on a windows computer.

So, I’ve spent around 3 hours going into every directory in the CVS tree, running the “cvs admin -kb” command on all binary files..

Followed by a couple of “cvs update -A”.

All this so I might checkout some files to eclipse on my stationary computer..

Google Image Labeler

Wednesday, January 17th, 2007

Google has released a very fun project where you and a randomly selected player are shown the same picture, and you have to enter words that describe it, if the words match you get “points”. The matching words are also used to improve the quality of the Google image search.

Click here to play (warning, it’s kind of addictive)

If you get teamed up with a decent player you will get like 700-900 points in a round (each round is 90 seconds).

in the worst case you might end up with a real numbskull, or a bot that only writes “Carcinoma” or other weird words on some of the pictures. I got the word “Carcinoma” on totally unrelated pictures at least 8 times, dunno what the hell that was all about.. probably bots, but they got the other guesses pretty right. weird.. Also the top score has been changing at random (3420600, 3309000, 3391600, 3409900) in a 10 minute timespan… But then again this project is still in beta

Aaanyway.. I finally got a good score (1100 points, ranked #1 at the time), thank you “other guest” :)
google image labeler