awediohead

Thanks for clarifying and confirming it's not something I need for now. It had been suggested to me as necessary previously, hence my question to double check. Cheers

--gpus=all --shm-size=256mb --mount type=tmpfs,target=/tmp/cache,tmpfs-size=100000000 --restart unless-stopped Can someone explain what the --gpus=all part does of the above Extra Parameters? When is it necessary / useful as its not in the default ?

I'm trying to set up Frigate as a docker with a static IP on Unraid BUT this is about my network and NOT specific to any particular docker, but a more general question about dockers on VLANs and custom networks. I'm also concerned that on pfsense's DHCP leases page I've got the same MAC address being associated with my server (as the IP address for my server) and the VLAN IP for my cameras: Where Unraid Intel eth0 (hostname) and 'singthesis-nas' (hostname) are 192.168.1.92 AND 10.0.44.200 respectively, with the Mac Address being 00:26:55:e1:xx:xx for both leases. Is this OK because I have a LAGG and VLANs between pfsense, my managed switch and two ports on Unraid or is it a problem waiting to happen? Put another way, the 10.0.44.200 is a DHCP automatically assigned IP because the DHCP range on that VLAN is 10.0.44.200 to 10.0.44.254, so it's just handed out the first IP in the range. On the Docker instance if I don't specify a static IP of 10.0.44.210 for Frigate it defaults to 10.0.44.2 which is not within the DHCP lease pool, so I'm worried that's going to cause problems? If I try to use --mac-address and use one of the docker specific Mac Addresses in Extra Parameters I get an error when starting the docker - something about ipvlan versus macvlan? Bearing in mind that there are doubtless large holes in my understanding of networking I'm probably asking the wrong questions and possibly seeing a problem that isn't there! Hope someone can help! P.S. I've tried to open the docker console and use ifconfig but it just says 'not found'

Oh I have - dozens of times by now! The problem is that states: As the camera I need to get working is 4K the example given is pretty much useless and I've had to search elsewhere for answers. Still, I'm getting gradually more familiar with the terminology and broad principles - but as many others have said, the Frigate documentation assumes a hell of a lot of knowledge. Even so with a working iGPU at least I can trial and error the process and make a little progress. If the built in .config editor wasn't checking syntax I'd have given up long ago.

Update: Turns out my BIOS had somehow reset the GPU aspect of itself (or didn't retain settings) such that the Intel iGPU wasn't working, while the Nvidia 1050 was, though only for transcoding in Plex. So as soon as I added the hwaccel_args (any type) Frigate crashed because it was trying to find something that wasn't there. And obviously iHD or i965 weren't going to make any difference either! At least now I can plough through the Frigate docs in a somewhat meaningful way. Are there any extra parameters I should be adding above and beyond the defaults for the 10400's iGPU?

So I've literally spent days on this and I'm going slightly mad. I'd really appreciate some help as I'm just going round in circles. I have a single 4K Reolink E1 Outdoor PoE camera installed for now. With two Reolink 520A cameras waiting to be installed. I can get VLC to show a nice clear image via: rtsp://admin:[email protected]:8554/h264Preview_01_main this also works as: rtsp://admin:[email protected]:8554/h265Preview_01_main (where obviously I'm substituting 'PASSWORD' for the actual password - which does not have any special characters!) but the same URL entered into Frigate just generates a series of errors: In the Log files I get things like: 2024-03-22 11:27:25.320784970 [rtsp @ 0x24c9e40] method DESCRIBE failed: 404 Not Found ERROR : e1frontdoor: Unable to read frames from ffmpeg process. 2024-03-22 11:27:25.426632987 [2024-03-22 11:27:25] frigate.video ERROR : e1frontdoor: ffmpeg process is not running. exiting capture thread... ERROR : Ffmpeg process crashed unexpectedly for e1frontdoor In the VAINFO section I get: VAINFO OUTPUT Return Code: 3 Process Error: error: XDG_RUNTIME_DIR is invalid or not set in the environment. error: can't connect to X server! libva info: VA-API version 1.17.0 libva info: User environment variable requested driver 'iHD' libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/iHD_drv_video.so libva info: Found init function __vaDriverInit_1_17 DRM_IOCTL_I915_GEM_APERTURE failed: Invalid argument Assuming 131072kB available aperture size. May lead to reduced performance or incorrect rendering. get chip id failed: -1 [22] param: 4, val: 0 libva error: /usr/lib/x86_64-linux-gnu/dri/iHD_drv_video.so init failed libva info: va_openDriver() returns 18 vaInitialize failed with error code 18 (invalid parameter),exit Changing from iHD to i965 makes no difference I've started from scratch so many times now I think there's something even more fundamental going on. I'd really appreciate some help figuring out what that is? My current hardware is a z490 +i510400 CPU so the onboard iGPU should be working in parallel with the USB Coral I have - which is at least being recognised in both Unraid and the Frigate logs. Hope someone can help. Thanks

So are you saying that - at the camera level - where I have (since just changing back to preset-vaapi) cameras: e1frontdoor: enabled: true ffmpeg: inputs: - path: rtsp://10.0.44.2:8554/e1frontdoor hwaccel_args: preset-vaapi roles: - record - path: rtsp://10.0.44.2:8554/e1frontdoor_sub roles: - detect detect: width: 640 height: 360 fps: 7 objects: track: - person you just have cameras: e1frontdoor: enabled: true ffmpeg: hwaccel_args: preset-vaapi detect: width: 640 height: 360 fps: 7 objects: track: - person or even less than this? Thanks a million

I do not have Portainer installed - in fact I had no idea you could install Portainer on Unraid ! While I will investigate Portainer further in future, but just now it seems like another steep learning curve when I'm already a bit overwhelmed with variables - and Frigate should work without Portainer right? I found this Reddit thread helpful as my three cameras are all Reolink - though I only have one wired up and installed right now. Here's my error logs: 2024-03-19 13:52:52.409885798 [INFO] Preparing Frigate... 2024-03-19 13:52:52.420791575 [INFO] Starting Frigate... 2024-03-19 13:52:53.723144268 [2024-03-19 13:52:53] frigate.app INFO : Starting Frigate (0.13.2-6476f8a) 2024-03-19 13:52:53.754498105 [2024-03-19 13:52:53] peewee_migrate.logs INFO : Starting migrations 2024-03-19 13:52:53.761128965 [2024-03-19 13:52:53] peewee_migrate.logs INFO : There is nothing to migrate 2024-03-19 13:52:53.767830216 [2024-03-19 13:52:53] frigate.app INFO : Recording process started: 584 2024-03-19 13:52:53.770085464 [2024-03-19 13:52:53] frigate.app INFO : go2rtc process pid: 89 2024-03-19 13:52:53.793262819 [2024-03-19 13:52:53] frigate.app INFO : Output process started: 595 2024-03-19 13:52:53.804977095 [2024-03-19 13:52:53] frigate.app INFO : Camera processor started for e1frontdoor: 603 2024-03-19 13:52:53.812644825 [2024-03-19 13:52:53] frigate.app INFO : Capture process started for e1frontdoor: 605 2024-03-19 13:52:53.836449259 [2024-03-19 13:52:53] frigate.video ERROR : e1frontdoor: Unable to read frames from ffmpeg process. 2024-03-19 13:52:53.836532504 [2024-03-19 13:52:53] frigate.video ERROR : e1frontdoor: ffmpeg process is not running. exiting capture thread... 2024-03-19 13:52:56.409451884 [2024-03-19 13:52:53] detector.coral INFO : Starting detection process: 593 2024-03-19 13:52:56.413665756 [2024-03-19 13:52:53] frigate.detectors.plugins.edgetpu_tfl INFO : Attempting to load TPU as usb 2024-03-19 13:52:56.413962714 [2024-03-19 13:52:56] frigate.detectors.plugins.edgetpu_tfl INFO : TPU found 2024-03-19 13:53:13.842880802 [2024-03-19 13:53:13] watchdog.e1frontdoor ERROR : Ffmpeg process crashed unexpectedly for e1frontdoor. 2024-03-19 13:53:13.843266663 [2024-03-19 13:53:13] watchdog.e1frontdoor ERROR : The following ffmpeg logs include the last 100 lines prior to exit. 2024-03-19 13:53:13.843625848 [2024-03-19 13:53:13] ffmpeg.e1frontdoor.detect ERROR : DRM_IOCTL_I915_GEM_APERTURE failed: Invalid argument 2024-03-19 13:53:13.843850708 [2024-03-19 13:53:13] ffmpeg.e1frontdoor.detect ERROR : Assuming 131072kB available aperture size. 2024-03-19 13:53:13.844088541 [2024-03-19 13:53:13] ffmpeg.e1frontdoor.detect ERROR : May lead to reduced performance or incorrect rendering. 2024-03-19 13:53:13.844255022 [2024-03-19 13:53:13] ffmpeg.e1frontdoor.detect ERROR : get chip id failed: -1 [22] 2024-03-19 13:53:13.844452915 [2024-03-19 13:53:13] ffmpeg.e1frontdoor.detect ERROR : param: 4, val: 0 2024-03-19 13:53:13.844676862 [2024-03-19 13:53:13] ffmpeg.e1frontdoor.detect ERROR : [AVHWDeviceContext @ 0x560671d1d540] libva: /usr/lib/x86_64-linux-gnu/dri/iHD_drv_video.so init failed 2024-03-19 13:53:13.844904200 [2024-03-19 13:53:13] ffmpeg.e1frontdoor.detect ERROR : [AVHWDeviceContext @ 0x560671d1d540] Failed to initialise VAAPI connection: 18 (invalid parameter). 2024-03-19 13:53:13.845005861 [2024-03-19 13:53:13] ffmpeg.e1frontdoor.detect ERROR : Device creation failed: -5. 2024-03-19 13:53:13.845178704 [2024-03-19 13:53:13] ffmpeg.e1frontdoor.detect ERROR : Failed to set value '/dev/dri/renderD128' for option 'qsv_device': Input/output error 2024-03-19 13:53:13.845337806 [2024-03-19 13:53:13] ffmpeg.e1frontdoor.detect ERROR : Error parsing global options: Input/output error 2024-03-19 13:53:13.864092082 [2024-03-19 13:53:13] frigate.video ERROR : e1frontdoor: Unable to read frames from ffmpeg process. 2024-03-19 13:53:13.864222820 [2024-03-19 13:53:13] frigate.video ERROR : e1frontdoor: ffmpeg process is not running. exiting capture thread... and here's my config.yml mqtt: enabled: False detectors: coral: type: edgetpu device: usb ffmpeg: hwaccel_args: preset-intel-qsv-h264 input_args: preset-rtsp-restream output_args: record: preset-record-generic-audio-copy go2rtc: streams: e1frontdoor: - rtsp://admin:<password>@10.0.44.101:554/h264Preview_01_main e1frontdoor_sub: - rtsp://admin:<password>@10.0.44.101:554/h264Preview_01_sub #oldcam: # - rtsp://<username>:<password>@<camera_ip>:554/h264Preview_01_main #oldcam_sub: #- rtsp://<username>:<password>@<camera_ip>:554/h264Preview_01_sub cameras: e1frontdoor: enabled: true ffmpeg: inputs: - path: rtsp://10.0.44.2:8554/e1frontdoor hwaccel_args: preset-intel-qsv-h264 roles: - record - path: rtsp://10.0.44.2:8554/e1frontdoor_sub roles: - detect detect: width: 640 height: 360 fps: 7 objects: track: - person I've tried substituing VAAPI for QSV and the generic vs restream for hwaccel and input arguments respectively. I also tried changing the iHD to i965 without any benefit. Hope someone can see what I'm doing wrong? Thanks

So having sent back a dud USB Coral TPU, I got a replacement and it's being seen on Bus 002 Device 003 Port 2-1 as. ID 18d1:9302 Google Inc. where the previous one never changed from the generic. Since I last posted on this thread, I've also changed my Unraid set up from being Ryzen based to using a z490 + i5 10400 Being at the, "doing what Ed (of SpaceInvaderOne fame) tells me to do, but being otherwise fairly clueless", end of the Unraid user spectrum and feeling pretty confused trying to make sense of the Frigate documentation, and having searched in vain for Unraid specific tutorials, written or video based . . . . 1) Can someone please tell me what order to work through the setting up of the configuration file? 2) Is there anything specific to Unraid that doesn't apply to Docker on, for e.g. Debian, I should be aware of? Knowing this will help me extrapolate info from tutorials for other systems. 3) I have a 10th gen (i5 10400) CPU with an iGPU - do I need or should I be making this available to Frigate in terms of hardware acceleration to work in partnership with the Google Coral TPU. I think this is so the TPU does object detection, while the iGPU does everything else more efficiently than the CPU? 4) Is it best practice to save and restart Frigate after making a change / addition to the config or to get all sections of the config set up as best I can before saving and testing? i.e. how interdependent are the various sections? Would the absence of a particular section cause it to crash or simply not work optimally? Thanks

Thanks - very useful real world info. So having just returned a dud USB coral TPU - would you recommend trying another of the USB type? I'm wondering because it's only an extra £10 on top of the USB unit's price for the dual TPU (https://amzn.eu/d/0ab2ffc) - though I assume I'd also need to buy some kind of adapter to fit a regular motherboard M.2 slot. That said I don't use any M.2 SSDs or NVME drives on my server (mostly for ease of access as my 2.5" SSDs are in front loading bays) so I'm pretty sure I should have enough PCIe lanes. The counter argument is that I also really like that the USB model would be very flexible - i.e. if I can't get it to work on Unraid (my Ryzen system has no iGPU and my GTX 1050 does transcoding on Plex) I could set it up on a spare 10th Gen Intel build that does have an iGPU and use that as a dedicated CCTV PC / NVR - without having to completely dismantle my Unraid server to make the swap ! Any other pros and cons to consider - assuming this is specifically for use with Frigate?

Thanks Bruceflix - accepting it's a dud probably explains why nothing else I've done has worked, assuming Frigate is looking for something that isn't there when the docker starts because that's what's in the config.yml ! I'll start a return and see what happens with the replacement. Cheers

I won't claim to have read every single post in the 43 pages of this thread, but I've read an awful lot of it, watched a bunch of videos and read (if not understood) quite a bit of https://coral.ai/docs/ and the troubleshooting section for the TPU (so I've tried multiple USB cables for e.g.) . . . and I'm still utterly clueless (and not a little overwhelmed!) as to where to begin with Frigate on Unraid specifically. The posts and videos I've read or watched aren't using Unraid and I just can't join the dots - doubtless because of fundamental gaps in my understanding of Unraid! So 1) Do I need to follow the instructions to Install the Edge TPU runtime on Unraid's console before I do anything else? 2) Do I then need to install the PyCoral library? 3) My USB coral never seems to initialise and change from 1a6e:089a Global Unichip Corp to Google Inc. - whats the quickest way of checking whether it's a factory dud? I assume I'll be plugging it into a Linux distro - I have a POP OS install on a PC, would that be close enough to being Debian/Ubuntu based to give me accurate feedback as to whether or not the TPU is functional? Any other quicker ways to determine this since I'd prefer to put most of my time and energy into figuring out how to get this working on Unraid! 4) I thought I had created a very basic config.yml file just to get Frigate started so I could modify the config from within Frigate as is recommended - but with the USB Coral TPU plugged in starting Frigate up and looking at the logs, the logs window flickers for a fraction of a second and then disappears far too quickly to read anything. That said am I right in thinking I need to be seeing the device as Google Inc in lsusb before I can hope to launch Frigate successfully? I appreciate I have a ton to learn to get things working, but I just can't seem to find anything that assumes my kinda level of ignorance!

I have much the same question - though with a Ryzen 3600 on x570. I'm really unclear as to what powertop can usefully/safely do for AMD CPU's? What it can't do and where to draw the line? The OP states it is useable, and warns that only getting C1, C2 and C3 is normal for new Intel and all AMD CPUs, however the guidance that follows assumes a fairly deep familiarity with Linux CLI that's way over my head. It's clear that I can't use autotune safely because of having a JMicron PCIe SATA card and the 'tunables' are 90% all 'Bad' apart from a dozen or so 'Goods' at the bottom of a very long list. No idea where to start! I also have a query about how you'd go about doing what's suggested in the OP without the config file editor which is now deprecated ? After messing about with BIOS setting for a few days at least I now have my system going to sleep when not in active use and responding to WOL to wake up again - this is pretty crude in terms of power efficiency but a big step in the right direction for me. I just went through every setting I could find in the ASUS x570 BIOS relating to C states and energy efficiency and turned it on or off - even something minor such as the onboard LEDs makes a huge difference to this stability IME. I'll research this more but maybe someone can chime in on the difference between the apparently problematic PCIe SATA cards mentioned in the OP (like my JMicron) and getting a PCIe SAS HBA card - I assume the latter does not have an onboard SATA controller chip and would therefore respond better to "instruction" from the OS directly in terms of power management? Probably wishful thinking but it'd be great if someone with the know how could do a guide for AMD power efficiency.

That all seems to have gone OK thanks. Only slight quirk is that it's complaining about eth1 being disconnected as if it's expecting a bond or LAG between eth0 and eth1 when I haven't set one up - the other two ports are marked as 'shutdown inactive' I think it's probably 'remembering' something from my previous half baked attempts as it says, "Interface Ethernet Port 1 is down. Check Cable!" in red on the bottom border of the page to the right of "Array Started" in green. Is there a way to clear this? The obvious restarts etc. haven't changed it. One other quick question at a tangent - if docker containers are running is it OK to just switch docker services off in Docker Settings, or should I be stopping them under the Docker Tab before stopping the service - or does it make no odds? Just want to know what best practice is.

Thanks very much! I feel I can go ahead now with a lot more confidence that I'm not about to do something dumb (again!) I think I'll keep it simple and just swap the NICs to start off with. I should be able to add the LAG afterwards once the unraid box is back where it lives next to the switch and router. And yes it is complicated - I have reams of notes and spreadsheets about how the switch, router and unraid are theoretically going to work together. Makes my head hurt, but kind of fun in a wirdly masochistic way cheers