Since then there have been several different spin offs and IMHO, better ones out there. For the last several years I’ve been running DD-WRT on my Linksys WRT54G (The same one that I originally put OpenWRT on, fried, and then fixed). I found myself need to extend my wireless network, but not in the normal since of wireless. I need to turn a switch into a wireless client so I could have a hard wired connection on the other end.

I started searching for a new router, one that would last me a while, and more importantly. One I could hack. The short of a very long process for me, I settled on a ASUS RT-N16. This thing is loaded, 32MB of storage, 128MB of ram, and480 Mhz processor. It supports B, G and N. Pretty nifty. I got it, fired it up. And spent just enough time in the Asus firmware to load up DD-WRT.

I took my old Linksys and converted it into a wireless client. So no, it is just one huge switch. What I have is the wireless portion of it is just like the wireless in a laptop, and it connects to my router. Then all the ports on the back are just like a normal switch. I can now plug my desktop computers in my office up to it, and they are all on my network. It’s a pretty slick setup.

The connection between the Linksys router (now switch) and the Asus router is a bit flaky at times. And this is because of a multitude of reasons. The biggest being the number of walls the signal has to travel through. But I am working on fixing that one (moving the Asus router out of the bedroom one day). The other problem is the DD-WRT firmware on the Asus router. The router requires a beta release of DD-WRT running the 2.6 kernel. Older versions used the 2.4 kernel. But to support a wider ranger or routers (including this Asus) they had to use the 2.6 Kernel. There are still some bugs, and some things are quite working like they are suppose to, but that is why it is a beta release.

Overall, I am very happy with everything. As long as I can resist the temptation of tinkering with them. The connection remains really stable.

MySQL allows you set to set the character-set and collation to utf-8. To my understanding this will allow you to store UTF-8 data in your database. This makes it really nice when you have users copying and pasting data into your CMS. It will store the special double-quotes, trademark symbols, and pretty much anything else you can throw at it. No big deal to backup up, right? Well I did an export within PHPMyAdmin and using the mysqldump utility. Neither one worked, and I was pissed. And actually so was the client because I didn’t catch it before we launched the site. So last night I was frantically searching Google as fast as I could (notice I wasn’t searching the web, I was searching Google – that should make an interesting blog post for another time). And I eventually found my answer. It had to do with the mediums inbetween the export and import to support utf-8 characters – Great.

So without having to use a 3rd party app, or write some code, I was able to figure out how to do it with myqldump. Instead of using the greater than symbol to redirect the output into a file, mysqldump has a switch -r to use to redirect output into a file. They say to use it on Windows machines, so I kind of felt dirty for using in the linux command line :/ But I did, and guess what, an import using virtualmin on the production server and everything worked, the data came across cleanly.

From the brief reading that I did, using the power of output redirection at the unix command line can break UTF-8 encoding. I guess it kind of makes sense if the server or terminal you are using doesn’t support it. But it isn’t something I would really think about.

Just for those wondering what I ended up doing. I normally do a mysqldump like

mysqldump -u user_name -p database_name > mydump.sql

This time around, i just had to change it to

mysqldump -u user_name -p database_name -r mydump.sql

So simple I could throw up.

While it is hard to trust google, and what they are monitoring while you would be using their OS, I welcome the change. The OS market is was developed when there was no internet , at least not how we know it today. Think of the Palm’s Pre and how superior it is over the iPhone b/c Palm centering the OS around the web. Andriod has a lot of the same feature. And now the Operating System. We no longer use one machine to manage our data. Hell, I use two machines just at work. No including the data and process that I work on in a server.

You can read more at Google Official Blog Post.

I can’t wait to play with it.

There are a lot of blogs and articles out there about SVN and NFS. You can search google and easily find a lot of information about them. But I never found anybody that felt like they found a definitive answer.

In my case I kept getting an error about not being able to find and available lock, the error looks something like svn: Can't get exclusive lock on file '/repo/db/write-lock': No locks available It turns out it is actually pretty easy to fix. And for those that don’t want to read further, all you have to do is add the nolock option to your nfs mount. My fstab entry now looks like this192.168.1.154:/volume1/svn /media/svn nfs rw,nfsvers=3,nolock 0 0. That’s all I use, and now my svn is working over nfs, safely I might add.

For those looking for a bit more explanation (which is what I couldn’t find). SVN dev’s prefer you not to use the Berkeley DB over NFS. So make sure you are running a new enough SVN so it uses the FSFS backend not the BDB. The next, it is my understanding that SVN uses file locks to protect the files during checkout and checkin, to prevent them from being over written. NFS uses NLM locking and it is used so that all clients of that nfs export no about the file lock. Apparently subversion is compatible with this locking mechanism. Adding the nolock option to the mount point as the client uses a more basic locking mechanism that only provides protection only form programs running on that machine.

What does all this mean, it means I am able to mount my svn share on a nfs export, but only that one machine can really access this export to guarantee consistency in our repository. You do lose some of the benefits of it being over an NFS share, you couldn’t do any type of load balancing over multiple SVN frontend servers, but it does allow our NAS to hold our repository and puts me one step closer to consolidating our storage.

While I didn’t provide a new answer, I am providing the reasoning behind the answer as I can see it. Something I couldn’t find out there on the Internet.

I’m really torn about what to use for my blogging software. I’ve gone through many different ones, and ended up on b2evo. I like, I really do, but it is falling behind times, themes are not being developed for it, and I don’t really have the time to write the things I want for it.

At the time I picked b2evo, there were some fundamental things I didn’t like with WordPress. But those things have changed, and I am really thinking about moving to it. Which sucks, but I’ve started to use my blog, and I have a good handful of blog post I really don’t want to use. There are some conversion utilities out there, but I’ve read mixed things about it. I don’t care too much about the links, I just want the thing to work.

So no point really to do this post, just thinking out loud.

Linux defiantly doesn’t have a lack of choice when it comes to file systems to choose form. I’ve use most of the them for my main file systems, currently I am running ext2 on /boot ext3 on /home and /, and using ext4 for my portage tree and distfiles on /usr/portage. In the past I have use reiserfs, reiser v4, jfx and xfs. They have all had their pros and cons, and I’ve have used them all on my /usr/portage. Interestingly enough, the best experience of them all is ext4. The others would become fragmented over time, since I am a bit anal about running the latest software I normally sync my tree almost daily. Most of the time every other day.so there are a lot of re-writes and small files scattered all throughout that partition. I have nothing but great performance using ext4. So much so, with the release of 2.6.28 I am thinking of moving my / partition to it.

This would give me /boot as ext2, /home as ext3 and everything else as ext4. I’ve read good things about it, it seems to handle large and small files alike with ease. I am actually pretty excited with the though of moving to it.

Hopefully my experiences will continue to be on the good side of things as I move to ext4!