Iceman24

Thanks. I will give it a go, just need to decide on SSD.

I have an unRAID build already, using HDD's as array and SSD as cache, but this 2nd one I'm doing has more simple needs and less storage capacity, so I was wondering if there was a reason this wouldn't be ideal. My thoughts were to put a 1-2TB SSD as the only array drive in lieu of 1-2 HDD's as array. I don't need much storage capacity, so SSD's are cheap enough that price wise it'd be around the same anyways. I would use SSD as the main drive for everything. The 1 HDD, only other drive in system I would use as a backup drive using Unassigned plugin. It could be external, but I want to keep it inside the computer. I would send backups to it from the SSD and any other computer(s) on the network, so it really would just be a backup drive, which is why I don't feel the need to mirror it with another HDD. SSD's are quite reliable, plus I'd be backing the data up locally to the HDD regularly, as well as the cloud, plus another SSD like that would really jump the price up. Any reason why this isn't the best method? I thought about the HDD as the array drive and just put everything on the cache SSD, but then why bother making it array drive if I really just store backups on it?

Should the docker image be on the cache anyways?

How are you paying $50 a year? I see $100 if paid annually.

Has there been any progress on getting this to work? I was going to try it since I could self-host, but I wouldn't be willing to do a VM just for it.

If that Docker is setup to only run on HTTP, the scheme needs to be HTTP. You can still force SSL in the SSL tab, but the scheme must be whatever the Docker is setup to use. Also, I would recommend blocking out your domain. Nobody needs to know that to help anyways.

You're welcome. Make sure this area of your config.php looks like this. 'trusted_domains' => array ( 0 => 'unRAIDIP:NextcloudPort', 1 => 'nextcloud.server.com', ), 'trusted_proxies' => array('unRAIDIP'), 'forwarded_for_headers' => array('HTTP_X_FORWARDED_FOR'), 'dbtype' => 'mysql', 'version' => '18.0.0.10', 'overwrite.cli.url' => 'https://nextcloud.server.com', 'overwritehost' => 'nextcloud.server.com', 'overwriteprotocol' => 'https',

I'm not in a position to do anything other than say check my post about 10 back with how to fix this. Have to add line in "default" to file.

Interesting. I never knew anyone included "php7" in the command and it worked for me without it.

How exactly did you type it?

Thanks, that did work. Odd thing though is that when I did so, it removed the ovewritehost from the file entirely and moved the overwrite.cli.url all the way to the bottom, but kept it. So it seems the overwritehost was the only one needing commenting or removal, as it was removed upon reboot and the other one was kept, but moved to bottom of file.

I had issues with the same was able to eventually figure it out. You need to console into the Nextcloud docker and run the below command, changing the end of it which perhaps you already did to get the BigINT conversion done. So "db:add-missing-indices" after the "occ" instead of what I have here. For the HSTS, add the line "add_header Strict-Transport-Security "max-age=15768000; includeSubDomains; preload;";" in the "default" file in the "\nextcloud\nginx\site-confs" folder. I put it above the line that has "SAMEORIGIN" in it, not all that far from the top of the file.

I thought updating apps through their built-in updater was a no-no on Docker apps, that you're only supposed to update them via Docker. Is it different for Nextcloud?

With my new install on this app on an i7 caliber machine, NVMe drive, is about 800Mbps down/250-300Mbps up on larger sized file a good speed? I was using web interface to test. For my usage, which would probably be mostly smaller files, super fast speeds is not a requirement, but I wanted to ensure I wasn't missing out on what I should be getting. I haven't tweaked any performance settings and this is on a reverse proxy.

Is this the redirect line and how it would look if I commented it? /**'overwrite.cli.url' => 'https://nextcloud.server.com', *\/

Not that it's a problem, but after changing config.php file to work with my reverse proxy, I can no longer access Nextcloud via the local IP address:port. It redirects me to the reverse proxy address, which works fine, but I have no other app that does this and if I had situation where that was the only way I could access it or just needed to that way for some reason, I would prefer to be able to, like I can for other apps. Why does this happen?

So far, my new install works with reverse proxy. I was confused seeing different little ways people were setting it up, but I used this video from SpaceInvaderOne as guide for editing only the config.php file. It worked straight away after that.

Okay, thanks. Does this mean the redirect will only work on the non-SSL port? I would've figured the cert issue wouldn't have been an issue if it redirected to where I told it to go.

Is there a way to add a proxy host, but not allow remote access from outside the network, just internal?

What is the proper way to update Radarr? From Wiki, it mentions not to update outside of the actual Docker itself, so I turned off Automatic updates within settings, but in this following screenshot, it tells me not to use 'develop' branch, but I used default settings to install it, so why would it be set to that then? i'm confused. What should I make the following settings?

I understand that 2 different services can't use the same port (on the same device). The firewall and unRAID server are 2 different devices, but I realize that connecting externally when they both have the same 443 port that there could possibly be a conflict and there is, but only when I have NPM trying to forward me somewhere for unknown domain used. Something about the firewall being on the same port had it redirecting me there even though I have no rule setup to allow outside access to firewall GUI. No big deal though, I had the port different like I mentioned, so I switched firewall port back. I just wanted to understand the technical reason for why that would happen. Errors for using SSL port for unknown domain on NPM are just like when you visit site with invalid certificate. FIrst, "NET::ERR_CERT_AUTHORITY_INVALID, then "ERR_HTTP2_PROTOCOL_ERROR" once I click to allow me to access site anyways. When I looked at certificate, it just a dummy Nginx Proxy Manager one.

Port 80 needs to be open in your firewall and it needs to direct into NPM, whether NPM uses 1880/80/whatever. You just need to setup firewall rule that allows external port 80 access into NPM.

I've noticed an issue that I mostly resolved, but I wanted some clarity on what was happening so that I understand it. Things were fine until I changed my firewall GUI port back to 443 from a custom port I had been using the whole time. I realized that after I closed off remote management access to my firewall, that I didn't really need the custom port # I had been using, so for simplicity, I put it back to the default of 443. Then I was testing my public IP in the address bar and noticed that if I go to default ports (80/443) with just my IP, I was taken straight to the firewall login page! I definitely didn't want that. I put my firewall GUI back to a custom port #, problem solved. I was then being redirected to the specified webpage that I told NPM to do when trying to access a proxy host on NPM that didn't exist. Problem solved, although I want to understand why this happened. Can you not have firewall GUI and NPM using same port without this conflict? Also, the redirect only happen if I use port 80. If I use port 443, it gives certificate error. Certificate is a dummy one from NPM, but even clicking through it, you just get error. It doesn't redirect to webpage I told NPM to use. How do I get that working properly on port 443? Thanks.

For this and any Docker, if the app has a shutdown feature within the app itself, should I do that before updating the Docker from unRAID?

Got it, thanks. Follow up question. Being that I can force SSL for connecting to apps that don't have SSL by using it on NPM, is it redundant to enable SSL for apps that actually do support it? If I just turned off SSL for all apps and relied on NPM for it, it'd be simpler and consistent for everything. Would that make sense to do or am I missing something?