jbrodriguez

?

I'll check for anything unusual in the code.

Glad to know it worked !

That's not fair, it shouldn't be doing that I'll take a look.

Yes, that seems to be the case. What unRAID version are you running ? Whenever you get a chance, can you $ killall controlr $ cd /usr/local/emhttp/plugins/controlr $ ./controlr It should start Api service on port 2382 and on http/https depending on how you set it up on your server.

I reproduced the crash while changing any setting. Also found out why it won't spin (on the Detail screen, but it works on the main screen I'll publish an ota release shortly.

Ashe/wgstarks, what iOS version are you on ? I'm testing on 10.3.3 and I can't reproduce the crashes yet I understand you were able to add the server. If you just refresh the server, without changing any setting, does it work ?

If you refresh the server (pull down), does it work ?

So, did you revert back to 2.10.2 ? Is that what you mean ?

v3.0.0 has been published on Apple & Google stores ! This release brings a new storage architecture, that makes the app easier to maintain and add features. You need to re-add your servers. These additional changes are included: - Configurable poke vs ping to check if server is online - Improved Settings and Add Servers screens - Used OS native fonts everywhere - Improved auto refresh behavior - Fixed power down/reboot issues - Fixed issue with scrolling marquee after orientation change - Other bug fixes and improvements Also, swipe to delete server is once again available by default. Just FYI, Poke checks that the port you specify when adding the server is accepting connections. Ping sends an ICMP request to the server. This should help with people that for some reason doesn't have ping available for their server(s). You need to have the plugin (2.3.0) installed, for optimal operation. Finally, I was going to implement ControlR as a notification agent, but held it back. It's relatively easy to do on iOS, but more difficult on Android, since there are different ways to do it and some providers/libraries haven't caught up yet with the current sanctioned way (FCM). So it will probably be a while. It also needs a new notification agent template that would need to be shipped directly with unRAID, as far as I can tell. I will look now into showing ups stats. Let me know about any issues you may find with this release.

Thanks for the report @wgstarks ! Yes, the powerdown/reboot issues are app related, I've fixed them in my dev environment. A new app release should be out soon™

v2.3.0 has been released 2017-10-04 - 2.3.0 - Add support for unRAID 6.4.x (rc9f), using Letsencrypt generated certificate by default (an alternate can be specified in settings) - Add endpoint for future ControlR app releases If you use unRAID's https, the plugin will now reuse the letsencrypt certificate (instead of generating a self-signed one). This means you will no longer have the ugly warning when accessing the plugin's webpage. If you want to use your own certificates (self-signed or LE generated by yourself), just specify the folder where the cert resides in the plugin settings page. The cert name must be certificate_bundle.pem (the same as unRAID's). This release supports 6.4.0-rc9f (and rc8q most likely), as well as stable (6.3.5). It should work on other unRAID versions, but they're not actively supported.

If you have movies movies/avatar/avatar.mkv movies/avatar/avatar.sub movies/avatar/poster.jpg movies/it/it.mkv movies/it/it.sub movies/it/poster.jpg and you select the movies share in the scatter page, it would move the 'avatar' and 'it' folders whole to the target disk. Probably 'avatar' would end in disk1 and 'it' in disk2 (as an example), depending on available space but each folder is copied whole.

Any plans for it in the future, if it's even possible? Not sure if it makes sense, to be honest. Will look into it.

No, it works at the disk level (/mnt/disk*), not at a user shares level (/mnt/user), so it doesn't know about allocation method nor split levels.

'nobody' is present by default in linux, so it shouldn't be an issue. Ye thats my thoughts too 'root' arguably has full access to the entire filesystem, so it should be able to read/copy any file and also delete them, if doing a move operation. I run unBALANCE as nobody and have run it with warnings, but this is because I knew the issues were about files/folder permissions not being exactly as expected from Docker Safe New Perms, not about owner/groups issues. The warning to run Docker Safe New Perms is general, since there are potentially many different scenarios.

Thanks for the info @AnnabellaRenee87. The issue is definitely that unBALANCE is displaying an incorrect representation (TB, GB) of the disk byte size. I will do some more checks and release a patch.

Thanks, one more thing, if you please. In /usr/local/emhttp/state/disks.ini, what's the fsSize for disk2 ? At the end of the day, it's a matter of javascript vs php code/math functions. But then, there's also this https://forums.lime-technology.com/topic/60415-odd-size-reporting/

Yeah, that's odd. Could you send me the output of df --block-size=1 /mnt/disk2 ?

I received two bug reports about not being able to power down a server with v2.10.2 and I was finally able to reproduce it. I've fixed it in my dev environment. The next version is a major internal overhaul (data storage architecture changed completely), which means when it's released you'll need to re-add you server(s) Sorry beforehand for the inconvenience, but a migration step would be expensive in terms of resources. Finally, I hope to make it available early next month at the latest.

Thanks for the kind comment ! From memory, it doesn't line up selected items with disks they appear on. It just lists selected items on one column and drives they appear in another column, so there's no 'horizontal correlation'. Not sure if that's what you're seeing. But I'll take a look anyway. I'll eventually add a select/deselect all for the TO column. Doing it one by one tends to get annoying

Yes, I will eventually do that Although, I can live without it, not a pressing matter

Oh ok. On 6.3.x, the app's Power Off sends a shutdown command to the server, which internally stops the array first and then powers it down. I didn't test your scenario, where you stopped the array first. I'll look into it.

The core idea is to allow unRAID's nginx instance to serve as reverse proxy to dockers and plugins, so you don't need an additional nginx (via a docker or such). Currently, /etc/nginx/nginx.conf has an include at the bottom include /etc/nginx/conf.d/*.conf; You could specify a nzbget.conf here and create a server block with a location to proxy 'http(s)://<ip|host>/nzbget' to the nzbget docker, for example. The issue is that the generated emhttp-servers.conf sets its main server block as a 'default_server', so it catches all urls directed at the server. Any other server block is essentially ignored, so nzbget.conf's server/location wouldn't work. The request is to have a include /etc/nginx/conf.d/locations/*.conf; as the last line of emhttp-servers.conf's server block. This way nzbget.conf would need just a location block and it would be picked by the main config file. The integrity of emhttp-servers.conf would still be enforced by nginx_test_conf. Hope you can accommodate this. Any other idea/suggestion would be welcome.

The message is a warning because if rsync can't transfer a folder/file due to a permission issue, it will throw an error, which will cause unBALANCE to interrupt the operation, and you end up with a partial transfer. If your files have read access for 'other', you should be able to complete the transfer, but any file/folder that deviates from this, will interrupt the operation. I'm not allowing unBALANCE to be run as root (and it's been like that since I converted it to a plugin), but I can't remember why