Before we get started you should read and follow my previous post because I borrow all of the config settings from Ubuntu’s Solr 1.4 packages.  The default config settings from the Ubuntu maintainers is still a decent starting point with Solr 3.4.  Once finished you can safely remove the old Solr 1.4 package if you want to.

With a working Solr 1.4 installation in place, we can get started on getting Solr 3.4 running.  You can change some of the following paths if you want, just remember to change them in all of the appropriate places.  Everything you’re about to see should be done as the root user.

Create some required paths

mkdir /usr/local/share/solr3
mkdir /usr/local/etc/solr3
mkdir -p /usr/local/lib/solr3/data

Next, re-own the data dir to the proper user

chown -R tomcat6.tomcat6 /usr/local/lib/solr3/data

Download the latest version of Solr

You can get the latest version of Solr from http://lucene.apache.org/solr/ and extract the files into root’s home directory.

wget http://mirrors.axint.net/apache//lucene/solr/<version>/apache-solr-<version>.tgz
tar zxvf apache-solr-<version>tgz

Extract the war Solr war file

Extract the Solr war file into a location.  You may need to install the unzip utility with apt-get install unzip.

cd /usr/local/share/solr3 

unzip /root/apache-solr-<version>/dist/apache-solr-<version>.war

Install additional libs

There are a few other libs included with the Solr distribution.  You can install anything else you need, I specifically need to have the dataimporthandler add ons.

cp /root/apache-solr-3.4.0/dist/apache-solr-dataimporthandler-* WEB-INF/lib/

Configure Multicore

If you want to have multicore enabled you’ll need to perform the following actions.  The rest of this post assumes you have copied this file and will require you to make some changes to support multicore.  I’ve marked steps that can be skipped if you also wish to skip the multicore functionality.

Copy in the multicore config file:

cp /root/apache-solr-3.4.0/example/multicore/solr.xml .

You should now edit the solr.xml file at this point, doing the following:

• Set persistent to true
• Remove entries for core0 and core1

Next, change the ownership and permissions so that tomcat is able to modify this file when needed

chown tomcat6.tomcat6 /usr/local/share/solr3
chown tomcat6.tomcat6 /usr/local/share/solr3/solr.xml

Copy existing config files

This is where we’re going to borrow some files from Ubuntu’s Solr package maintainer.

cd /usr/local/etc/solr3
cp -av /etc/solr/* .

Because we simply copied the config files we need to modify them to fit our new environment.  Change the following in the solr-tomcat.xml file:

• Change docBase to /usr/local/share/solr3
• Change Environment value to /usr/local/share/solr3

Also edit tomcat.policy file changing:

• Modify all entries referencing solr to point to appropriate /usr/local location

Change the following in conf/solrconfig.xml:

• Change <dataDir> to /usr/local/lib/solr3/data

If you are using multicore and you followed the Solr 1.4 multicore post you’ll have a conftemplate directory and you’ll need make changes to conftemplate/solrconfig.xml

• Change <dataDir> to /usr/local/lib/solr3/data/CORENAME

Create symlinks

Here we’ll create some symlinks to support the way Ubuntu packages Solr.  This is necessary because we copied Ubuntu’s config files and those files reference a few locations.  Creating the symlinks also allows us to continue using the scripts created in the previous post with minimal modifications.

• cd /usr/local/share/solr3
• ln -s /usr/local/etc/solr3/conf
• ln -s /usr/local/etc/solr3/ /etc/solr3
• ln -s /usr/local/lib/solr3 /var/lib/solr3

Enable/Start the new Solr instance

We can now enable our new Solr 3.4 instance in tomcat by doing the following:

cd /etc/tomcat6/Catalina/localhost
ln -s /usr/local/etc/solr3/solr-tomcat.xml solr3.xml

Note that the name of the symlink is important as it will define where we find this instance (/solr vs /solr3).  At this point you can create a new core.  I’ve provided the updated scripts here.

I’m now releasing version 0.3.0. This version brings a few changes but most notably the Linux client is now ready. The next release will be coming shortly and will focus on making the client the more robust about how it deals with network disconnects.

You can read more about the 0.3.0 release at https://github.com/dustinrue/Dencoder/wiki

This thread was the key I needed to get it all figured out – http://www.mail-archive.com/linux-raid@vger.kernel.org/msg10661.html

After issuing dmsetup remove followed by the device name shown in dmsetup table I was off and running.

Two, the /etc/default/squid.conf SQUID_MAXFD is not listened to either, you must set this in /etc/squid/squid.conf instead.

https://bugs.launchpad.net/ubuntu/+source/squid/+bug/580590

The community edition still has InnoDB built in as an available storage engine but you can’t buy support from Oracle.

http://www.mysql.com/products/
http://palominodb.com/blog/2010/11/04/oracle-not-removing-innodb

Expect is a way to programmatically work with a normally interactive process. Using expect you can write a script that telnets into a system and then issues commands based on what it “sees.” Here is the script I used, with some important values removed, to automate the process of updating a number of devices.

#!/usr/bin/expect
set timeout 300
spawn telnet 192.168.1.1
expect "login: "
send "root\n"
expect "Password: "
send "tehmagicphrase\n"
expect "# "
send "cd /tmp \n"
expect "# "
send "tftp -g -r config.ini 192.168.1.159\n"
expect "# "
send "config.sh import config.ini\n"
expect "# "
send "tftp -g -r firmware.img 192.168.1.159\n"
expect "# "
send "firmware_upgrade /tmp/firmware.img 1\n"
expect EOF

The above script was saved into a file called pushConfig.expect and set as executable using ‘chmod +x pushConfig.expect’. To run the script, I powered on the device and waited for it to be ready, once ready I issued ./pushConfig.expect to start the update process.

Using expect is fairly straightforward. The most difficult part is ensuring you correctly tell expect what to look for before sending the next command. In the script above I do the following:

set timeout 300

This tells expect to wait at least 5 minutes for matching text before continuing to the next send command. What this means, is if I tell it to send some data it’ll wait up to 5 minutes to see what is in the expect line after the send. In the case of my script the firmware upgrade could take quite a bit of time and I didn’t want it to timeout so I set the value fairly high.

The next line tells expect to start a telnet session to a remote machine and then to wait until it sees:

login: 

Once it sees that it sends the username. The script continues like this until it sees EOF. At this point expect knows that the process is now complete and it exits.

By using an expect script I was able to simply power on the hardware device and wait for it to boot. Once booted I ran the script. This saved me and a co-worker a lot of time while pushing custom configurations and upgrading the firmware on a number of devices.

Expect is capable of a lot more than I used in my example and can react differently based on what it receives back from the interactive process or even loop over a series of commands. To learn more about expect try ‘man expect’ or search your favorite search engine.

…the future for KDE technologies on mobile platforms certainly looks bright.”

I really beg to differ.

(via OSNews)

The only “problem” I had was that my Linux system would show up with an icon that looked like an Apple display. I wanted to appear in my Finder window as a little bit closer to reality. By default OS X is aware of all of Apple’s products and can display the appropriate icon for any product it finds. In my setup I’m happy with my Linux system appearing as an older style of PowerMac. With a bit of effort you can cause OS X to show any custom made icon you want.

This blog post explains the process and this comment describes the various device types you can specify.

Labeling a partition is simple, just use e2label /dev/ “” where is the device and partition you wish to label and is what you want to label the partition with. For example, if you added a second hard drive to your system and created one large partition, you would probably label the drive with the following command

e2label /dev/sdb1 "NewDrive"

To mount this drive on boot, edit your /etc/fstab file as root with your favorite editor. Add a new line at the bottom with, based on the example above, the following:

LABEL=NEWDRIVE /mountpoint defaults 0 0

You can use the other entries in the file as a template. Use ‘man fstab’ and ‘man e2label’ to learn more about how to use e2label.

The other interesting thing, as always, are the comments on the Slashdot post. I just can’t help but feel the majority of tech savvy people simply don’t understand what the iPad is and why it will be a success. According to a large number of the tech savvy, people want a device they can tinker with, install whatever they choose on and is open. Anything else is evil and you’re dumb if you don’t agree. What they tend to forget is there is a large market of people who just don’t care about openness. They just want a device that works without the need to tinker.

My wife isn’t someone who wants to tinker. She wants results. As a tech savvy person myself, it took me a number of years before I fully understood why turning on the TV, stereo, selecting the right input, turning on the DVD player and selecting the right remote for each device in order to watch a movie simply wasn’t acceptable to her. It had to be simpler. The entire setup was so simple in my mind and I had complete control over the system, it all seemed perfect to me. Getting everything turned on so we could watch a movie together seemed like a small accomplishment each and every time. To her it was a complete chore. The answer, in the end, was a Logitech Harmony remote. One button turns on the needed equipment, automatically selects the correct input and the one remote then correctly controls each device without further work. Turning it all off is equally easy. Today I’d never give up that remote because of how it simplified the entire process.

My point is that eventually, getting things done becomes more important than openness or the ability to tinker. I’m more than willing to give some of that up for a more pleasurable experience. I think Android based devices, including phones and tablets, will be successful, but in the area of tablets the iPad will lead because it provides solutions with the least amount of resistance.