Iceman24

Members
  • Posts

    107
  • Joined

  • Last visited

Posts posted by Iceman24

  1. Just now, Djoss said:

    The Force SSL option redirects users connecting to the non-secure HTTP port (80) to the secure HTTPs port (443).  Since NPM is mainly used to access services over the Internet, having an encrypted connection is always better.

    I understand the SSL is better, but when you can force it when connecting to the non-SSL port, is there a point of using the SSL port? If I never opened the SSL port on my firewall or connected to NPM using the SSL port, but I ultimately get redirected there, is opening SSL port/connecting directly to it have any real benefit?

  2. Is there a point to using the SSL port at all for, whether connecting to NPM externally or internally, if you use the force SSL option which effectively gives you SSL anyways? Thinking about it, it seems as if it's double SSL getting to where you're going. I don't know if that requires anything to work harder, but after discovering I could force SSL on apps that didn't even offer it, I wondered what the point of using SSL for any specific apps was in the first place? It would seem to only deal with SSL for NPM and not any other apps you use it to connect with as the most efficient and simplest option. Do I have this right?

  3. I'm stumbling across this issue just before I setup my new build for unRAID. I have 2 970 EVO PLUS 500GB's I was going to use a cache pool using BTRFS. Guess I won't bother with that, can't hardware RAID them with current hardware, not going to get anything else to do that. I want redundancy with them, that's why I bought 2. Should I just get different SSD's? If so, which ones are good with no issues?

  4. I was all set to install my USB 2 flash drive on a internal header that gives me a USB 2 port (it itself is plugged into USB 2 header), but then reading

    Quote

    Avoid using front panel USB ports in favor of ports available directly on the motherboard I/O panel.

    on the Getting Started page gives me second thoughts, despite hearing that putting it inside was safer as it won't stick out. I know a 6" header adapter isn't a front port, but that is how a front port connects, so what's truly the best way to do it? I want to do it that way. Thanks.

  5. On 6/26/2019 at 4:58 AM, _jonte said:

    Time to get a couple of new drives.

    What type of WD are best for this?

    I'd like to be able to put them back together if a get a failed drive again.

    The My Book Desktop, and Elements Desktop, are about the same price.

    They're similar, all very shuckable. Videos galore on doing so.

  6. , There are no different between add-in card and mainboard build-in. Pls well planning and research.
    Thanks. Are there in particular NVMe drives that are free of issues? I've also read in one place that performance is barely different compared to SATA due to sequential reads/writes not really being used, but mostly random, which has similar speeds. SSD's have become cheap enough that I wanted the latest and greatest for minor cost difference, plus more clean install in case.
  7. For example
     
     
    Lol. Funny you picked that one. I was last poster in that thread. I figured that it may have had to do with using an add-in card for connecting M.2 drives. Motherboard I'm looking at has them built-in. I read similar post as well that referenced only issues with M.2 that used add-in card instead of built-in slot.
  8. I'm about ready to order parts for my unRAID build, but had a question regarding PCIe lanes usage with dual M.2 NVMe cache drives. I'm looking at the Supermicro X11SCH-F motherboard which has 2 M.2 NVMe ports with PCIe 3.0 x4 for each slot. No mentioned sharing of anything. I won't have more than 2 USB drives plugged in and 3 HDD's at max most likely. Perhaps a GPU card for passthrough at a later date, but not planned at this time.

     

    From what I read, SB has its own PCIe lanes, so that seems okay from that perspective. The only thing I'm really wondering as I am figuring the 2 M.2 slots support full speed with each having a drive in them, is the upstream of the C246 chipset to the CPU is only x4, not the full x24 the chipset offers. Does this matter for M.2 cache drives? They will be RAID 1 if that matters.

  9. Depends on the specific GPU. Many detect what monitor is attached and only offer resolutions supported by that monitor, so if you have nothing attached, you may not get any output. There are dummy plugs you can get to fake the GPU into running properly.
     
    You will still need to use a remote access program hosted on the VM like nomachine, RDP, teamviewer, logmein, gotomypc, or similar.
    Would the P630 integrated graphics in the Xeon E-2176G work for passthrough to a VM? What about if I wanted to pass it through to Plex docker for transcoding (which I don't often do), could it not be passed through to both simultaneously?
  10. I'm having trouble figuring out exactly how this works. First, unsure if I'll use VM's on my new unRAID server I'll be building soon, but I would like to be able to if I decide to.

     

    I understand that you can passthrough a GPU that you would then connect a monitor to physically on the server and use a VM that way, but can you passthrough a GPU to a VM so that it can utilize it so that it runs better, looks better, smoother framerate, if you only connect to the VM over the network, instead of being physically plugged into the passthrough'ed GPU's video output port?

  11. 14 hours ago, Djoss said:

    You don't really need to have a split DNS setup.  You should be able to access the service from both the outside and inside using the DNS name pointing to your internet IP.  Since your are using pfSense, you can just enable NAT reflection to make it work.

     

    This is how I'm setup with OPNsense (similar to pfSense) and it works wonderfully.

    Thanks, but from prior research NAT Reflection isn't the recommended way to handle such routing. It's recommended to leave it off and use Split DNS, so I am determined to keep it configured that way. Even if nothing else worked, I'd rather work around it be using port 443 internally/externally for NPM and workaround that port being unavailable for unRAID GUI, which I don't access remotely anyways.

  12. Other than switching unRAID's ports from 80/443 to something else and switching NPM from 1880/18443 to 80/443, is there a better way to setup access to an app that I access from both inside and outside my network that no app I'd access it from allows you to set local and remote ports to use depending on whether you're connected to your WiFI? My issue is that for something like Nextcloud which I'd self host, setting it up via the reverse proxy (which I'd much rather do than just using port forwarding, which I'm trying to get away from for all my apps and utilize NPM instead), causes the issue of NPM by default uses different internal ports than it does external, so it's not practical to reconnect to my server with the Nextcloud app whether I'm inside my network or away every time I need to use it. I have Split DNS setup, but it doesn't change the port, so this doesn't work. I don't want to use Pure NAT or whatever variant of NAT that isn't really the right way to go about it.

     

    I've been trying to find out the best way, which might be with HAProxy with pfSense, which is what my router is, but it's setup seems complicated and I don't really know for sure if it does work like that based on research I've done anyways. If I swapped ports with unRAID like I first mentioned, this seems like it was solve issue the easiest, but I'm not sure if that would cause issues or annoyances that aren't worth the hassle. I just want to be able to use the same server IP/Address whether inside or outside my network. Some apps allow this with some of the apps on my unRAID, but not Nextcloud, and not a couple others.

    • Upvote 1
  13. Did the issue logging into the UniFI controller ever get resolved? I read about it in this topic, but never saw a resolution. I'm encountering the same thing when connecting through the proxy, whether externally or internally. The UniFi app works fine though using the proxy. I've read elsewhere that people have applied some sort of "referrer" code, but I tried sticking it in the advanced section and mostly just knocked the proxy host offline. Some lines of code that apparently resolved it was the following.

     

    	location / {\n\
    	proxy_set_header Referer \"\";\n\
    	proxy_pass https://localhost:8443;\n\
            proxy_set_header X-Real-IP $remote_addr;\n\
            proxy_set_header X-Forward-For $proxy_add_x_forwarded_for;\n\
            proxy_set_header Host $host;\n\
            proxy_http_version 1.1;\n\
            proxy_set_header Upgrade \$http_upgrade;\n\
            proxy_set_header Connection \"upgrade\";\n\
    	}\n\

     

    It does work with Firefox, but not with Chrome.

  14. Can the external and/or internal ports be changed for this app? Whether I wanted to use different ports when coming in from outside, instead of 80 and 443, or possibly use 80 and 443 internally, which I realize would make me change it for unRAID, but I wanted to know if doing either would cause issues with anything? Also, is port 80 needed anyways? I use only SSL.