Jorgen

Thank you, thank you, thank you!!! I tried adding a mouse and keyboard earlier but couldn't get it to work. I incorrectly assumed this was because of vt-d missing, but your post prompted me to have another crack at it. For some reason I had selected "USB Mode = 3.0 XHCI" when I first created the VM. Once I spotted that in the XML and changed it back to "2.0 EHCI" things started working again. Thanks again for your help, and for all the awesome guides your are producing!

Hi all, I've shot myself in the foot and was hoping someone on this thread could help me out. I have a 10.12.5 VM that was created with the clover method and was running fine, but during troubleshooting of an unrelated problem I set my Ethernet interface to "inactive" via the network preference pane. Face palm! I lost connection to the VM of course, because it has no means of communicating with the outside world anymore. And now I can't connect to it to restore the network preference. I had splashtop and apple's screen sharing working, but can't connect to either anymore (not surprisingly) NoVNC and other VNC clients can connect while the VM boots up, but as soon as the apple logo appears I lose mouse/keyboard functionality. I can still see the boot progressing up until the login screen, but that's as far as it gets. Without any input method I can't login. My CPU doesn't support vt-d, so I don't think I can pass a physical mouse/keyboard to the VM? I'm open for any suggestions at this point. Here are some options I've considered or tried: 1. Mount the HFS+ partition of the vdisk image and manually edit/remove the network plists. I've only managed to mount it read only so far, but I'm also a bit worried about these fits as I'm not at all confident that I'm doing the right thing. 2. Add another virtual network interface in the XML. Maybe there's a way to trick OS X to think I have a wifi or other interface that can give me network access. I have no idea where to start on this though, or if it's even possible. 3. Create a new VM and restore my user account from the latest time machine backup. This will take hours and I won't learn anything new, so it's my last resort. Also, I had problems restoring from time machine when I first created the now broken VM. 4. Buy a vt-d capable CPU and pass through physical mouse, keyboard and GPU. This is actually my preferred solution but will take me some time to get the funds together and then waiting for the order to ship. Not sure I can live without the VM for that long. Sorry for the long post. If anyone have any ideas on best way to proceed please share. I'm posting here because there aren't many Mac-VM-on-unRAID threads around. Happy to start a new thread if more appropriate. And thanks to gridrunner for making all this possible in the first place!

Actually, I can still login with the default Admin/Password even after I've removed the Admin user. Guess the CLI password change is still needed.

Excellent! Thanks for testing. Sent from my iPhone using Tapatalk

I thought the default admin account was recreated on update of the docker though, is that not the case? Sent from my iPhone using Tapatalk

Exactly Sent from my iPhone using Tapatalk

Ok, so my symlink problem was due to my own misunderstanding of how to create them. I know you tried to tell me this Djoss, but it took me a while to understand that you can create a symlink pointing to any path, whether that path actually resolves or not. For other's benefit, here's one way to achieve this. Assuming you store your media on unRAID like this: /mnt/user/Media/Movies/moviefile.mkv And you have the following HandBrake folder mappings: /storage = /mnt/user/Media /watch = /mnt/user/Media/Handbrake-hotfolders/watch Then from the unRAID console, to create a symlink in the watch folder that resolves from within the docker container: ln -s "/storage/Movies/moviefile.mkv" "/mnt/user/Media/Handbrake-hotfolders/watch/moviefile.mkv" Now, from an unRAID perspective this symlink is most likely broken, because /storage does not exist. But from within the Handbrake container, the link resovles fine thanks to the folder mappings. And it works great. So that's the last of my problems, all good now. Thanks again Djoss, really like this docker!

I've configured my (custom) preset to retain ALL audio tracks (and subtitles), but I haven't actually tested the results. If I get a chance I'll have a play with it this weekend. Sent from my iPhone using Tapatalk

Doh! Thanks guys, don't know how I managed to miss those two things, seems so obvious in hindsight! So deletion of files in watch is now working perfectly for me, thanks again! I think it has also "fixed" my other problem with multiple files in watch on docker start, as each file deletion seems to trigger a rescan of the watch folder... Will have to read up on symlink creation. My ultimate goal is for NZBGet to store the HD file to my normal Media share, and also put a symlink to it in the Handbrake watch folder (but only whenthe movie is listed on my friends trakt.tv watchlist...) If anyone knows how to create a symlink and change the folder part to allow for the docker folder mappings, please speak out! Otherwise I'll report back if/when I solve it...

Hi Djoss, thanks for this! I went to town with the testing and unfortunately not everything passed. Quick summary: Files aren't deleted from watch folder after succesful conversion Multiple files in watch folder seems to cause problems triggering detections of all files. It stops after each conversion, requiring another file to be added (or deleted) to convert the next file (but only one) Symlinks only works if the target file is in the same directory (watch folder). In hindsight I'm not sure this was a great idea, since /storage and /watch are mapped to different folders inside the docker the link is probably broken from Handbrake's perspective? Or maybe i'm not creating the symlink correctly (linux noob...) Details of test cases and results here: https://docs.google.com/spreadsheets/d/11voL9X4c9d9IL18NGx1m8SAcM_smPhoxvYPixJzpEJU/pubhtml Let me know if something is unclear and I'll try to explain it better.

At the risk of sounding greedy here, but would you also consider adding an option to automatically delete files from the watchfolder once they are succesfully processed? Thanks!

Agree with dlandon, of the handbrake dockers I've tested this is the easiest to configure and works the best out of the box. For example, this one lets me add files to the watch folder while the docker is stopped, and when started will happily chew through the backlog. Neither of the previous dockers let me do this. (I am using CA User Scripts to schedule all handbrake processing overnight) I've found a small bug though: File names containing spaces will make the watch folder skip the file. To reproduce: 1. Add a file to the watch folder called "A file with spaces.mp4" 2. Check logs. There will be a series of skipping file messages, one for each of "A", "file", "with", "spaces.mp4" And while I have your attention, does the watch folder work with symlinks instead of actual files? If not, consider this a feature request

Oops, sorry about that. Wrote it from memory since copy-paste from ssh/mc to safari on an iPad proved a little bit too challenging... Sent from my iPhone using Tapatalk

Hi Kizer, this is what I've done and it works for me: 1. From the HandBrake GUI, export your custom preset to a file and store it in your config folder. For example: appdata/HandBrake/mycustompreset.json 2. Edit the HANDBRAKE_COMMAND line in appdata/config/Handbrake.conf like this Default Config HANDBRAKE_COMMAND='HandBrakeCLI -i "$SRCDIR/$SUBDIR/$FILENAME" -o "$DESTDIR/$SUBDIR/$BASE.mp4" --preset "Very Fast 1080p30"' New config HANDBRAKE_COMMAND='HandBrakeCLI --import-preset-file /config/mycustompreset.json -i "$SRCDIR/$SUBDIR/$FILENAME" -o "$DESTDIR/$SUBDIR/$BASE.mp4" --preset "mycustompreset"' IF your preset have spaces in the name, you need to enclose the name in double quotes. More info here: https://handbrake.fr/docs/en/latest/cli/cli-guide.html You should be able to import the preset directly from the GUI version with --import-preset-gui but I never got that to work. Plus I wanted my custom preset to be stored outside the container anyway, in case I have to re-install in the future.

Minor bug: When you hover over the name of a script and click "edit name", the pop-up menu doesn't collapse automatically (like it does for the other options) If you click on "edit name" again in this state, code snippets are being added to the name box. Each click adds more code.

Latest update seem to have fixed this problem for me, thanks Coppit!

appdata/Handbrake is set to 99:100 for me (It's possible that I've set this manually during troubleshooting) The folder I have to delete to get it working again is: appdata/HandBrake/handbrake That one is recreated on docker start and seems to have correct settings. The PUID/GUID is set in the appdata/HandBrake/HandBrake.conf root@Tower:/mnt/cache/appdata# ls -l | grep HandBrake drwxrwxrwx 3 nobody users 117 May 15 22:22 HandBrake/ root@Tower:/mnt/cache/appdata# tree -plug /mnt/cache/appdata/HandBrake /mnt/cache/appdata/HandBrake ├── [-rw-r--r-- nobody users ] HandBrake.conf ├── [drwxr-xr-x nobody users ] handbrake │ ├── [-rw-r--r-- nobody users ] Activity.log.92 │ ├── [-rw-r--r-- nobody users ] ghb.pid.92 │ ├── [lrwxrwxrwx nobody users ] handbrake -> /config/handbrake │ ├── [-rw-r--r-- nobody users ] preferences.json │ └── [-rw-r--r-- nobody users ] presets.json └── [-rw-rw-rw- nobody users ] processed_files.dat I see that coppit is updating some files on github and it looks like the file that creates the error (/etc/my_init.d/start.sh) has been updated to remove the symlink to the handbrake config (which is part of what I'm deleteing) so maybe he's across the problem and solution already. I'll just wait patiently for a container update...

Apologies, I got that wrong. I can actually leave the file config/HandBrake.conf as it is after editing. To get the container working again I have to delete the whole directory config/handbrake then just start the container. The directory and files are recreated and it all works until next container restart (including scheduled appdata backups).

That's the exact same error log I got previously. I thought I had worked around it, but following this mornings appdata backup I'm in the same situation as you, container refuses to start. hmm, maybe it's the appdata backup that's causing problems?

Sure does. I think coppit outlined the changes to this watch folder in sparklyball's handbrake thread. This one is much more reliable, you can add new files at any time and they will be processed. With sparkly's I could only get it to process the first batch of added files. Sent from my iPhone using Tapatalk

Thanks for this container, @coppit! I swapped from sparklyballs as I could never get the hotfolder processing to work properly. I think I've found a bug though, is anyone able to confirm they're seeing the same behaviour as me? Clean install of container, only set the folder mappings Container stops automatically (expected) Edit user and group ID in config/HandBrake.conf Start container. Starts up fine. Stop container. Edit preset in config/HandBrake.conf Start Container. Error. Container won't start. Revert change to config/HandBrake.conf Start Container. Error. Container won't start. At this stage no matter what I do I can't get the container to start. To restore functionality I have to delete config/HandBrake.conf and start over from step 3. Funny thing is, if do the preset edit at step 3 together with the user/group edit, the container starts up fine. It's like you only have one shot at editing config/HandBrake.conf, and subsequent edits will brake the container.

ok so it looks like it does write lots of crap to stdout in console mode, i think the timestamps your seeing are triggered from the auto refresh for nzbget web interface. i have tweaked the image again and it now closes stdout, ive also created /data/dst as for some bizarre reason it doesnt create it, even though thats where the logs go, sounds like a bug to me but hey its now fixed for the docker image, so please pull down the latest and give it a whirl. Success! Logs are nice and clean again. Thanks for the quick fix, and thanks for putting so much effort into these dockers, it's very much appreciated! Sent from my iPhone using Tapatalk

Just updated today, and now the docker log is getting spammed with this: 2016-10-07 22:09:27,019 DEBG 'nzbget' stdout output: [1d[37m[44m8[24;80H(B[m[39;49m[37m[40m 2016-10-07 22:09:28,029 DEBG 'nzbget' stdout output: [1d[37m[44m9[24;80H(B[m[39;49m[37m[40m 2016-10-07 22:09:29,044 DEBG 'nzbget' stdout output: [1d[37m[44m40[24;80H(B[m[39;49m[37m[40m 2016-10-07 22:09:30,054 DEBG 'nzbget' stdout output: [1d[37m[44m1[24;80H(B[m[39;49m[37m[40m 2016-10-07 22:09:31,061 DEBG 'nzbget' stdout output: [1d[37m[44m2[24;80H(B[m[39;49m[37m[40m 2016-10-07 22:09:32,071 DEBG 'nzbget' stdout output: [1d[37m[44m3[24;80H(B[m[39;49m[37m[40m 2016-10-07 22:09:33,078 DEBG 'nzbget' stdout output: [1d[37m[44m4[24;80H(B[m[39;49m[37m[40m New entry every second. Is anyone else seeing this? I did change Setting/Logging/WriteLog within nzbget from 'Append' to 'Rotate', but no other changes made apart from this. I've reverted that config change but the logs are still getting spammed.

I finally got around to upgrade my BIOS to the 2003 version and tried switching on the virtualisation. I found the VT-x setting but there is no VT-d setting. Could be because my CPU does not support it!! i3-4130 does have VT-x but not VT-d Can I still do virtualisation in unRAID? What would be the restriction? Thanks Yes you can. I have this mobo with an even lower grade CPU (celeron) and I manage to run two VMs. One Linux and one Mac OS X. VT-x lets you do VMs. VT-d lets you pass through hardware to the VM, for example a dedicated GPU for gaming. So dependent on your needs, VT-x might suffice. If not, you need to upgrade your CPU to something that supports VT-d (the mobo does already)