Let me clarify LinuxRocks will not be down the whole weekend just a few hours as it copies. I am doing it this weekend. Its 500GB over a 500Mbit line so it should not take to long.

@omnipotens not that I care much but this what really sucks about non fault tolerant federated networks. Even though it's decentralized, the integrity of data is the instances burden. It should at least be moderately fault tolerant. Heck the so called centralized social networks all have redundancy. Why can't a decentralized network?

Anyway I'm glad for the people on linuxrocks that you are hosting the community and maintaining it. But I do feel for these kind of issues...

Edit: just to clarify when I said fault tolerant, I meant not only the data integrity but also availability

@joelfernandes agree it be nice if I can setup two servers to fail back and forth were one will mirror the other. I think I could get the database to mirror but the application with all the images not sure how to pull that off. I thought about running symcthing on the application side and setup the database to mirror. Then I could failover with dns. two issues is I haven't had time to test it and secondly syncthing is a heavy load itself. Still might mess with it if I get time.

Sign in to participate in the conversation

Linux Geeks doing what Linux Geeks do..