Munin has been monitoring all my servers for some time now but I have always used the unencrypted connection because there is not much secret data transfered anyway and the built-in ssl connection did not work right away.
Now I found a small guide to building an ssh-tunnel for munin. The solution is obvious, why didn’t I do that? Anyway, need to do some changes to my munin setup during the next days.
At last Skype has got Video support on Linux. Today the Skype Linux team announced a new beta with video support. I have installed it on my new Ubuntu machine and it worked right out of the box. Unfortunately my screen captures did not contain the video image so I cannot provide an image here.
Anyway, great feature - thanks to skype!
Last week I tried to use two virtual servers (based on Virtuozzo) to learn something about DRBD. As DRBD syncs a block debice I wanted to set up a loopback device and sync that across both servers.
Unfortunately all my tries to get the loopback device to work resulted in an error “no permission to look at /dev/loop#”.
Today I found the solution in the OpenVZ forum: there is no support for loopback devices in machines running on OpenVZ/Virtuozzo.
For a real long time I have been looking for a good backup solution for my notebook. A long time I had been using boxbackup to backup up to a fileserver at my employer. While being the sysadmins makes this a possible solution I didn’t like the idea of backing up all my personal data to a disk in the company.
Another solution was using reoback and a wireless harddisk at home - but this suffered from the vfat filesystem on the disk and bad throughput on wifi.
Some weeks ago I stumbled over remote storage systems and especially Amazons S3 (Simple Storage Service). Amazon is offering expandable diskspace on its cluster systems with a pay-per-use pricing. What does that mean? There are now setup and no monthly fees, you really only pay for what you use. All stored files are calculated on a gigabyte/month price and this is really broken down to a daily calculation. So if you store 1 gigabyte for only 3 days, they charge you only for 0.1 gigabytes-months. Added to this are fees for requests (directory listings, storing and retrieving files) and data transfer.
The problem with S3 is that it works as a webservice and so there is no straight-forward way to put files there.
The solution for me is called Jungledisk: Jungledisk offers a webdav access to data stored in amazon S3 and thus makes it accessible with standard system tools. Thanks to gnome-vfs (and I understand KDE has a similar systems) virtually every software can access the data. Added to this is the ability to schedule automatic backups with include/exclude lists and bandwidth limits.
What is missing? Right: Amazon is a big company but that does not force me to trust them when storing all my data. Jungledisk solves this by encrypting all data on the client system before uploading it to Amazon. That way neither Jungledisk (as a company) nor Amazon can open any of the files.
Setting up Jungledisk is easy. All you need to do is download the package, extract it to a folder of your choice and run “junglediskmonitor”. The software can be used free for a test phase of 30 days. I tried it for a day and then chose to buy it: The price tag ist 20$ for lifetime upgrades and of course you need to pay the monthly usage fees to amazon s3. So far I am paying between four and five dollars a month for the backup and have saved all my personal data to an encrypted external storage. When my notebook drive failed last week Jungledisk already saved a lot more than it has cost me so far.