Posted on in Desktop, Linux

Trying out Fedora 26: No way-land?

My relationship with Fedora is of the sort where every two years I look at screenshots and go ‘Nice, why am I not running this again?’ Then I install it and everything looks fine. Then bit by bit the seams start to show.

Fedora 26 was to have been my grand entry into all things Wayland. The laptop running Ubuntu 16.04 started to feel like it needed something fresh or maybe I was just bored. I went in with open eyes, knowing that the move from Xorg to Wayland would invalidate a whole swathe of hacks I have used since times immemorial. It was going to be painful but a) it’s the cost of progress, might as well bite the bullet now and b) something something profit? Profit being some unforeseen advantage to doing things the new way.

To cut to the punchline, Wayland-Gnome runs and is perfectly servceable and I applaud the Fedora team for being the vanguard of the future and all that. I’m just not convinced that I as a regular user want to be on that frontline.

(more…)

Posted on in Desktop, Linux

Why my udev backup script kept failing and how I fixed it

I wanted to use udev to sync my backups to an external usb drive but the rsync-based shell script kept mysteriously stopping before the job was done.

When I say ‘use udev for backup’ what I mean is to use udev’s discovery of me plugging in the offline backup drive to fire off the backup script. Plug it in and the synchronization starts automatically. When it’s done it will automatically unmount the drive and inform me of the outcome via email. Sometimes there would be several weeks of new backups equalling gigabytes of data and so it would be convenient to just plug it in and leave it to do it’s thing.

The script was not apparently crashing wildly, just at one point it is running and the next it is not.  htop shows it to be running. iotop shows that data are being transferred. And then poof, suddenly it’s gone, even though there were still tens of Gbs left to transfer.

Luckily, the explanation is simple. Udev can fire off scripts but they are supposed to be very shortlived. Syncing 40 Gb over a USB 2.0 connection takes over an hour. The script is killed by the system after a few minutes. I found this elegant solution in an opensuse forum post.

(more…)

Posted on in Desktop, Linux

Steam In-home Streaming lags with NetworkManager revisited: Replacing NetworkManager with systemd-networkd

Back in early 2015 I encountered an issue with NetworkManager: Even if you have ever only told it about your one single wireless network (or AP, access point) it insists on scanning the airwaves at regular intervals (about every two minutes) looking for a better connection. I thought that was dumb. This behaviour is known to the developer and is working-as-intended. There is no off-switch. I thought that was really, seriously dumb. I still do.

It was not an issue for me until I tried live streaming, specifically using Steam’s in-home streaming that allows you to play a game on a different computer to the one you’re sitting at and streaming the ‘playback’ to your current machine. Then NetworkManager cuts into the smooth streaming every two minutes, turning it into a screenshot slideshow instead for 3-5 seconds. If you’re playing a game that requires attention to not dying/crashing/clicking-the-wrong-button, this is a bad thing for a long time.

Now, my gaming desktop died shortly afterwards (note to self: Never buy another (cheap) Corsair PSU as long as you live) so it quickly became a non-issue. I recently got a new gaming computer, dual-booting Windows 10 and Ubuntu 16.04 17.04.  Coupled with a new 802.11ac wireless network  it seemed obvious to try to make in-home streaming work with the desktop-to-HTPC streaming, essentially making the HTPC a high-end Steam Link, rather than a low-end gaming computer in it’s own right. Running the streaming ‘server’ from Windows, everything works. Streaming from linux I was instantly reminded of the problems I had two years ago.

About half a year ago, commenter William Bernting kindly added some information to my original post suggesting how to make NetworkManager stop the scanning routine. I hope it works for whoever finds this information and I should probably have tested his solution before forging ahead with my own, more radical one. However, I have no patience with NetworkManager anymore, I do only connect to one network and I had the impetus last night so here is what works for me: Rip out NetworkManager, replace with systemd-networkd.

(more…)

Posted on in Desktop, Linux, Server

Chorus2 + Firefox 51 = Easier music streaming

One of many new features in Kodi 17 is the new default  web frontend, Chorus2. The old Kodi frontend was an extremely basic web remote that allowed you to start, pause and select media. Chorus2 does all that in much slicker package AND has a ‘local’ tab for local playback, i.e. streaming from the Kodi instance to whatever device you’re controlling it from.

(more…)

Posted on in Linux

Upgrading to Kodi 17: The unintuitive bits

  • Move the old .kodi profile to .kodi backup and start over. Judging from the changelog there have been a lot of changes to the media libraries so we might as well start afresh.
  • On Ubuntu you will need to apt-get dist-upgrade to get Kodi 17 rather than just upgrade because of new dependencies. This is assuming you’re working with the official PPA.
  • On Ubuntu you’ll need to manually install kodi-inputstream-adaptive  (as a package using apt-get, not an addon in Kodi) in order for DASH to work. DASH is used by the YouTube addon to allow for greater and more variable streaming quality.
  • On Ubuntu you may have to manually install kodi-peripheral-joystick (again as a package) to get your controller working.
  • Installing SuperRepo v17 allows for browsing the packages but either something’s amiss with the repository or I’m being a dumbass. Lots of ERROR: CCurlFile::Stat – Failed: Couldn’t connect to server(7) in the kodi.log, though. To download a package manually from Superrepo.org, e.g. Steam, click the button that displays the number of downloads, the click SuperRepo location and save the zip file. Then choose Install from zip file in the Add-on browser (the cardboard box icon in the Add-ons menu). Or maybe you can just add the Jarvis repo instead of Krypton and get away with it?
  • Favouriting an add-on directly in the Add-ons menu is not possible as the option does not appear in the context menu. Go to the My add-ons submenu and find the add-on in the proper category, e.g. Program add-ons for Steam, Video add-ons for YouTube etc. Context-clicking it there will provide the option to favourite it.
  • Getting a play queue in the music player?
Posted on in Desktop, Linux, Server

Running Kodi as a pure client on linux

When I got rid of the last Windows partition on any of my home computers, I thought I would finally give NFS a chance to replace CIFS as the reigning network file system in my house. To it’s credit, NFS took that chance and ran with it and it’s worked pretty much flawlessly ever since. Seeing how reliable it has been, it has become my primary means of accessing media located on my HTPC when not in front of/within earshot of the my TV/hifi but still on the local network.

That approach, however, has some serious shortcomings. I can’t see what I have and have not watched. I can’t see if I’m halfway through a movie – something that is depressingly common in this ADD world of media abundance. I have to move between file managers and video players, something that isn’t that practical when my 2-in-1 laptop is in a position that hides away the keyboard and I have to rely on the touch screen. And I find myself missing the inviting presentation of Kodi from my HTPC.

So recently the thought struck me: Could I not just set up a second Kodi instance on my linux laptop only as a client? Turn a full-fledged Kodi application into a dumb terminal? An… *makes sign of the cross* ‘app’? Only then I did I remember that, wait, wasn’t that sort of the selling point of Plex over Kodi? Apps aplenty? Streaming everywhere and anywhere? Well, yes. But it can be done with Kodi. Sort of. Here’s what I got.

(more…)

Posted on in Python

Squeezing the sponge: How to implement a buffering SMTP Handler in Python

…when he needs what you have gleaned, it is but squeezing you,
and, sponge, you shall be dry again.
– Hamlet

One of the first things I implement in most Python projects is some quick stream logging so as not to dot too many print statements across the code. At some point file logging usually  follows as a feature so it can run by itself and I can check up on it from time to time.

Recently I got ambitious and added email logging to poca, my podcast client project. Which isn’t difficult as the logging module has an SMTPHandler built-in that you just stick on your logger and now your logs are airborne. It’s almost xkcd-like. The way it works though is that an email is sent off instantly every time something is logged, just like stream logging only over port 25. This was not what I wanted. To be specific I was looking for something that

  • Would gather up log entries as my script runs it course
  • At the end of the script it would evaluate if there was anything worthwhile sending
  • If so, send it; if not, save, don’t send

The way to do, as suggested by the logging module author himself, is to subclass the BufferingHandler which does a fine job of soaking up the entries and keeping them until its time to release them. He has written an example of how to do this himself and put it on Github.

I had, however, a further requirement. When a podcast feed fails, it’s not necessarily cause for alarm or notifying the user. Maybe the server’s in maintenance mode for a few minutes, maybe builders unplugged the cable for an hour, maybe you did. You only really need email logging to tell you when a podcast feed consistently fails because then it’s probably either a software failure or the feed is offline, either of which requires user intervention. Vinay Sajip’s script doesn’t allow for that. Also, using unicode is easier if we keep a clear distinction between body and headers which Sajip’s approach seems to fudge a bit.

My requirements meant that I would have to not just buffer entries in memory but save them over time and only release them when there was a sufficient number (and of sufficient severity) to worry about.

(more…)