Well, my rpi isn’t backed up. The data that’s on it is all on github so all I would have to do is restore the OS. I have yet to compile anything from scratch as I use pre-compiled packages and install via 'apt-get'. However, for other, more dedicated servers that I run, I do back them up every single day, however again, I don’t bother with the OS. What I do backup (list based on a CentOS machine):

/home - where everyone's user accounts reside
/etc - because it contains a myriad of config files for the whole system
/var/lib/mysql - all the MySQL databases live there
/var/spool/cron - because I really don't remember all the various custom service tasks that are running
/root - because I tend to keep some important, system wide stuff there
/var/mail - users tend to lose their heads if suddenly their entire INBOX goes missing

Everything is done using 'rsync' in incremental hard-links mode. This way I can keep (on average) 2 weeks worth of data without it also taking up 14x the amount of physical storage; it only copies a file if it has changed. I have a single terabyte drive that has been the backup drive for 15 different servers, storing anywhere between 2 to 4 weeks worth of data (and it’s also backed up to a secondary machine). One of the servers takes up 175GB all on itself. If I didn’t do incremental hard-links, I’d be needing a larger backup drive just for that one machine.

Should the main hard drive in any of the servers die, all I have to do is swap it out for a new one and push the power button. PXE (pixie) takes over, sees that there’s no OS on the drive, and automatically re-installs everything for me from the source OS server on the network. 10 minutes later the server is up and running, fully configured, with a restore rsync running. That’s faster than many will take to finish a hot cup of coffee in the morning.

I realize this is a more advanced setup that will not apply to 95% of you here. But then, you did ask how we backup. :)

If I were to backup my rpi, that’s the same process I would take, with the exception that rpi does not support PXE, so I will still have to manually re-image the SD card and go through the setup process to get it back up and running, prior to doing any kind of restore operation on it. And since I use my rpi only for MM, there isn’t a whole lot in terms of what needs to be redone on it, other than a few git clone commands and restoring my config file. But since I run an identical setup at my office but on a regular unix server (for when I’m bored and want to do dev work), I just copy that file over and done.

One of these days I’ll map out how my internal network runs and “maintains” itself.

I haven’t forgotten, it’s just been two rather busy weeks. And I’m recovering from a failed server HD today/this weekend. Yay for backups … not so much fun when it’s terabytes of data. Anyway, I’ll get back this as soon as this mayhem stops, or at least subsides a bit.

@patm189, my config is in the post above yours. I actually have 3 calendars listed that way.

Guys, use proper formatting for your code! (edited to add: when posting on the forums - not in your config files)

Block code is done with ```on a line by itself, then your code, and closed with another ``` on a line by itself again. This is not the same as an apostrophe! It’s called a 'back tick' or 'accent grave' and can be found to the left of the numerical number 1 at the top on a US keyboard. Please use it! It makes reading your config files a lot easier.

On non-US keyboards, um, its location will vary. You can also press and hold your ALT key, and using the numerical pad, enter 9 followed by 6, then let go of the ALT key. Repeat that three times and you get the three back ticks necessary.

As @sullit mentioned, Github is a code repository, any code, or to be more precise, any text. It’s a way for a coder to keep revisions on their code, and able to publish it to the general public. The code lives on whatever development machine you are using, and you sync, or mirror it to Github. This allows a coder to work on their code locally, break it, fix it, change it, do whatever, and when they’re done, push it up to Github to store and eventually release to the public. This also allows a coder to roll back their code. If while working locally they broke something and can’t figure out what happened, they can pull code back out from the repository and either merge it back in, or just compare and see where things went haywire. It’s a beautiful thing.

And while it may have a bit of a steep learning curve, trust me when I say this, it will save your sanity.

There are a lot of different guides, both on Github as well as out in the wild about how to use it.
Overall: https://guides.github.com/
Short and sweet beginner’s guide: https://guides.github.com/activities/hello-world/

Those are configuration settings that are explained in the wiki. However, it’s worth noting that it only works if you have the 'x11-server-utils' package installed as you’ve discovered. On a stock Rasbian Jessie install, that package does get installed, but I suspect those with different distributions, or a Jessie-Lite install won’t have that. I’ll amend the wiki shortly. Thank you.

If you must use liquid to adhere the film to the monitor, I would recommend putting a small bead of silicone along the edges to seal the glass to the frame. Or, if you are removing the frame, you can seal all around the glass screen itself with a thin layer of silicone before you put the liquid on.

@leisi said in Please help:

@darrene said in Please help:

sudo apt-get update

After the sudo apt-get update again i get following error again:
(connection via SSH from Mac / my OS is Raspbian Jessie)

pi@magicmirror:~ $ cd MagicMirror
pi@magicmirror:~/MagicMirror $ npm start

I lost count on how many times this has come up already. You can not start the mirror using that command when you are remotely SSHd into the rpi. The rpi simply does not know what monitor to use, yours, or the one connected to it. You have to tell it that:

DISPLAY=:0 npm start

Use the correct command, and then figure out what the problems are.

The fading doesn’t make it less obvious. It happens because of the different sizes per module. Fade or not, you will see the other modules in the same region jump to adjust to the new sizing. Some days my calendar will have long entries, and it makes the clock above it move. So when I hide/show the calendar, the clock will jump back and forth. The regions being flexible is what allows for longer entries on the calendar to be seen, for example. Same with the news ticker as you’ve discovered. Having fixed regions can cause other problems as well.

But, this should be a start for you. As I mentioned, it’s crude, and hard coded. I much rather have a function you can pass details to but I will leave that as an exercise for you. :)