Setup MySQL 5.1 on Centros 5.3

Posted on April 22, 2009 at 4:05 pm in

Yes I’ve done it. Finally, with much pain and suffering I’ve setup mysql on a Centros 5.3 slicehost.com box. Took me about a 1 1/2 days to set everything up from Apache/MySQL 5.1/PHP as well as a few other pieces of software. The main problem is incompatibility between MySQL 5.1 and PHP.

Currently there are two blog posts out there with 2 different methods of setting up MySQL 5.1 and the rest of the LAMP software.

http://gravityfx.org/2009/04/08/how-to-install-php-529-andmysql-5133-on-centos-53-final/

http://consciencespeaks.blogspot.com/2009/01/setting-up-lamp-server.html

I was able to get the gravityfx.org version by using the remi repositories. The only difference / change I had to make was:


yum enablerepo=remi install mysql-server

and leaving out the mysql since it installed packages/software from the centros base that interfered with remi’s packages.

Anyways either way it’s not the default setting and the php installed was missing some important libraries for tools such as phpMyAdmin.

The main problem for all of the installations is that php is not compatible with the MySQL 5.1 libaries. It works fine with libmysqlclient 15 but not libmysqlclient 16 as is bundled with MySQL 5.1.

I would recommend anyone installing software on linux to just stick with the defaults as they are what works.


Hadoop: Buggy on Windows.

Posted on April 16, 2009 at 10:39 am in

So yeah running hadoop on a windows box has a few bugs / issues that will definitely frustrate you as a developer.

If you suddenly start getting strange errors from hadoop and can’t seem to fix them by either restarting hadoop with stop-all.sh/start-all.sh the next best thing is just to clean that sucker up.

How do you clean up the hadoop distributed filesystem?

1. bin/stop-all.sh
2. delete the temporary files/directories c:\tmp
3. bin/hadoop namenode -format
4. bin/start-all.sh
5. bin/hadoop dfs -put local_dir dfs_dir

Of course that means you just formated/removed all of the data you put into the DFS but if all your doing is testing and development it shouldn’t be a big deal to put that data right back.
Here’s an example of an error that’s nearly unfixable and will just pop up and ruin your day for no reason:

09/04/13 14:48:35 INFO dfs.DFSClient: org.apache.hadoop.ipc.RemoteException: jav
a.io.IOException: File /user/Adminestrator/input/better.PNG could only be replicated to 0 nodes, instead of 1

Bummer.


Top