If you ever thought why I should buy a Mac the answer is simple, Alfred ! Can’t think of my work day without it. Soon to be release V2 of Alfred will include a new feature called “Workflows” and as a Mega Supporter I have access to beta version of the app and let me tell you this, if you’re a Power Pack user, prepare for your mind to be blown !
Continue reading
Dashing – The exceptionally handsome dashboard framework
I love dashboards, or actually monitoring dashboards, especially small simple scripts giving you a nice view into what’s important for you.
In my time at Rackspace I have build a Bootsrap PHP polling monster showing a lot of details about Rackspace Cloud (metrics, graphs). This help me and other admins to have a easy way to look into our infrastructure. The main thing is, all of it is read only. However, as much as I love Bootstrap it’s not ideal solution for large TVs on the wall were data needs to be pushed and things are BIG. Here comes Dashing Continue reading
Can’t unlock System Preferences padlock
Another problem I had after upgrading to Mountain Lion was that every time I tried to open the padlock in System Preferences my password was not accepted, here’s a quick fix:
user@computer:$ sudo rm -r ~/Library/Caches/com.apple.systempreferences
then reboot your system and you should be good.
Java fix for Mountain Lion
After upgrading to OSX 10.8 i got plugin not installed error in chrome when trying to open kvm java console. To fix Java open your termianl and type:
user@computer:$ java -version
you should be prompted to install Java update.
Puppet load balancing with Cron
Running Puppet from a Cron is usually better idea than running it in agent mode. However you may overload Puppet master with requests. This happens when you have 100s of agents trying to pull catalogs from master at same time. Lets consider this setup:
600 nodes, agent runs from cron every 5 minutes, this means you are going to have 600 agents trying to get a catalog from the master at the same time. You are going to see some failures caused by time outs. I was looking into this problem and found some good examples of randomizing cron runs (link). This is all good but does not guarantee balanced runs equally spread in time. Here’s my go on this problem. Continue reading