• tar xzf CrashPlanPRO_2010-03-08_Linux.tgz (or latest file)
• cd CrashPlanPRO-install
• sudo ./install.sh
• finish install with all defaults
• /usr/local/crashplan/conf/my.service.xml and change "servicehost 127.0.0.1" to "servicehost 0.0.0.0" and reboot. (you need to change this back once you get it installed so the world can't make changes to your backups)
• make sure your firewall allows port TCP 4242 and 4243

Then you need to install CrashPlan Client on your local pc and make the following change:

• c:\Program Files\CrashPlan\conf\ui.properties change "#serviceHost=127.0.0.1" to "serviceHost=10.0.0.5" (10.0.0.5 is the Linux server ip)

Now you can connect using the client on your pc but making the changes on the Linux box.

(Thanks Schwer Log)

Quick intro: A friend recently had a duck take over his planter and use it as a nest for its eggs. So, he bought an IP camera.

The Foscam fi8905w outdoor wireless IP camera has some nice features, but it needs a lot of hand-holding to make it work. You can control it through a URL-based set of scripts. Here's the most useful examples:

Camera status: http://my.ip.or.url/get_status.cgi?user=username&pwd=password Camera settings: http://my.ip.or.url/get_camera_params.cgi?user=username&pwd=password Snapshot from camera: http://my.ip.or.url/snapshot.cgi?user=username&pwd=password Reboot: http://my.ip.or.url/reboot.cgi?user=username&pwd=password Param set example, changes camera mode to 'outdoors': http://username:password@my.ip.or.url/camera_control.cgi?param=3&value=2
You can set parameters and reboot the camera with commands detailed in this guide.

We ended up with a series of tools and scripts to make the camera more useful. First is a script that grabs a photo every minute and stores it in an archive.

#!/bin/sh wget -O /www/duckcam/downloading.tmp http://username:password@my.ip.or.url/snapshot.cgi >/dev/null 2>&1 mv /www/duckcam/latest.jpg /www/duckcam/archive/cam.`date +'%Y-%m-%d-%H%M%S'`.jpg mv /www/duckcam/downloading.tmp /www/duckcam/latest.jpg
Next, because the camera doesn't handle going from night to day very well, we had to write a script to handle that. But, the camera stops responding after the brightness has been set, so we wrote a script that reboots it once a day right before setting the brightness at daybreak. This seems to work, but it's far from ideal.

#!/bin/sh wget -O - "http://username:password@my.ip.or.url/reboot.cgi" >/dev/null 2>&1 sleep 60 wget -O - "http://username:password@my.ip.or.url/camera_control.cgi?param=3&value=2" >/dev/null 2>&1 sleep 20 wget -O - "http://username:password@my.ip.or.url/camera_control.cgi?param=1&value=48" >/dev/null 2>&1 sleep 20 wget -O - "http://username:password@my.ip.or.url/camera_control.cgi?param=1&value=32" >/dev/null 2>&1
We realized that we wouldn't be able to stream from the camera itself because the owner's internet connection wouldn't handle it, so we found a tool that takes the IP camera feed and rebroadcasts it. The tool is called webcamXP.

After all that we have a pretty good setup for watching the miracle of life. Even if the image is a little washed-out.

That's what I just did and here are the most valuable command line arguments for doing so:

tar -pcvzf www.tar.gz www/ mysqldump -u root -pPassword --all-databases > alldata.sql tar -pxvzf www.tar.gz mysql -u root -pPassword < alldata.sql

Of course, those commands are not all that I needed to do, but they are the ones that I will want in the future when I move to another web server. So, I guess you could say they're more like a personal note on how to transfer a directory without losing permissions and how to move an entire MySQL install.

grep -i 'mydomainbasename' /unzippedlogfiles/other_vhosts_access.* > mylog.log

Attacking mylog.log produced results that were much more in line with what I was expecting.

rm -rf

The -r means recursive, and the -f means force; which means it will ignore the fact that the directory is full.

I'll start by saying Samba is not as easy as it should be, even when it's fully set up and working, but it gets the job done, and SFTP gets so tedious after a while.

Here's what I had to do to get it working. First, install it:

apt-get install samba

Next, edit /etc/samba/smb.conf and add something like this:

[www] comment = My Shared Stuff path = /my/folder browseable = yes read only = no

(I added it at the bottom of the file so that I could find it quickly)

I reloaded samba at that point, but I'm not positive it's necessary:

service smbd reload

Then, and this is the part I had forgotten from setting it up last time, I added my user to the list of samba users:

smbpasswd -L -a username

Done. Now get back to work!

Edit:
Now that I have installed the GUI, just do it from there. Right click on the folder to share it then run this command…

sudo smbpasswd -a username

Linux permissions are tricky for a Windows guy like me. When I copy a file to a directory on a server I expect the file to inherit the permission of the containing folder on the server. Period. Why? Because that's what NTFS does; that's how it works! Linux it's a bit more complex.

I don't understand it enough to explain it, but here are some of the links and notes on what I've learned so far.

I activated ACL for the volume I was working on.

apt-get install acl

I had to edit /etc/fstab so that the / line read something like "/dev/sda1/ ext3acl,defaults,errors=remount-ro 0 1" then reboot.

This allowed me to set permissions and default permissions for each folder using ACL. Here's the command I used:

setfacl -R --set u::rwx,g::r-x,o::- /www/ setfacl -R -d --set u::rwx,g::r-x,o::- /www/ setfacl -R -m g::rwx /www/wordpress/wp-content/ setfacl -R -d -m g::rwx /www/wordpress/wp-content/

This pretty much took care of what permissions were allowed for each new object. But I wanted my new objects to have the same group as the parent, and I wanted it's subdirectories to inherit that group:

chmod g+s /www/wordpress/ cd /www/wordpress/ find . -type d -exec chmod g+s {} \;

That's about as good as I could make it. It's not the Windows way, but it will do.

EDIT: Having thought about it some more, I'm thinking that I didn't need to activate ACL for my purposes. So I'll probably end up restoring a snapshot that I took before I started messing with ACL. Still, it was a learning experience.

history -c

Over the course of a year I figured out that the iCal glitch happens every few months; iCal forgets that it has scheduled tasks and doesn't run them. Why does this happen? I have no idea.

I've finally gotten around to fixing it and figured I should create a cron job, but Apple says "cron is so 1980's. You should use launchd!"

Creating a launchd job is not as easy as cron, but the concepts are pretty simple. Basically, this is how you do it:

1. Create a plist, according to the specs, that describes your job. I found that using Apple's Property List Editor, included with the Apple Dev Tools, made the task much simpler and helped me avoid some syntax mistakes that I was sure to make otherwise.
2. Place the plist in the appropriate folder. In my case I put it in /Library/LaunchAgents (check the spec to determine the appropriate location according to the task you're trying to perform)
3. Load your job like this: launchctl load /path/to/your.plist

That was it. In the end I was able to get it to work by launching a bash script from launchd. In turn, the bash script launched my original AppleScript. It's kinda crazy and convoluted but it worked. AFP458.com has a great tutorial on launchd that helped me more than the apple document.

MAke sure you have a mail server installed…. something like postfix.

Edit /etc/aliases

added or edit an entry for root to forward email:

root: myname@gmail.com

Save the file.

Run the program 'newaliases' which will update the aliase database and test it.

Thanks, man!