[Support] spaceinvaderone - Shinobi Pro


Recommended Posts

4 hours ago, CanuckleHead said:

Hello,
I am in need of assistance trying to get all this working with a VLan.  I'm trying to follow the videos from @SpaceInvaderOne.  Here is my setup:

Wireless Router

router_LAN.png.3fd6787c372462b61e2b629fcc887183.png

Router_DHCP.thumb.png.d424a15cc841f35f05012b0794e2db46.png

 

I bought a tp-link TL-SG108PE managed switch.  This is how I've set it up:

router_IP_Settings.png.a89fb053521b62c2215b6b43063890d5.png

Router_VLAN.png.6cebc4c4ac754d89215c11d0a494424f.png

 

Finally, here are my Unraid settings.  Network

Unraid_Network_Settings.thumb.png.40a3dcacea2dd8477f44e2f3107de077.png

Docker

Unraid_Docker_Settings.thumb.png.678cf25a8a8198aba52b83d3cbde6d31.png

 

Finally, my Shinobi docker settings

Unraid_Shinobi_Settings.thumb.png.1ecd1b12856f152f52d168e165865cf4.png

 

From my Unraid tower, I'm able to ping the Shinobi IP of 192.168.50.10.  However, when I try to launch the GUI, it just spins and finally states that it can't reach the page.  

Prior to trying to set up the managed switch & VLAN, I was able to get to the gui when everything was connected/set-up using the normal network.  I know in the videos SpaceInvaderOne did mention that he was going to have another video showing the VLAN setup but I can't find it.  Also, I'm not currently using PFSense.  I don't know if my current setup (Asus wireless router) supports the VLAN.

 

Any assistance with this would be greatly appreciated.  I'm trying to set this up in a secure setup.

Thanks

 

I pretty sure your router doesn't support vlans

Link to comment
11 hours ago, CanuckleHead said:

Hello,
I am in need of assistance trying to get all this working with a VLan.  I'm trying to follow the videos from @SpaceInvaderOne.  Here is my setup:

Wireless Router

router_LAN.png.3fd6787c372462b61e2b629fcc887183.png

Router_DHCP.thumb.png.d424a15cc841f35f05012b0794e2db46.png

 

I bought a tp-link TL-SG108PE managed switch.  This is how I've set it up:

router_IP_Settings.png.a89fb053521b62c2215b6b43063890d5.png

Router_VLAN.png.6cebc4c4ac754d89215c11d0a494424f.png

 

Finally, here are my Unraid settings.  Network

Unraid_Network_Settings.thumb.png.40a3dcacea2dd8477f44e2f3107de077.png

Docker

Unraid_Docker_Settings.thumb.png.678cf25a8a8198aba52b83d3cbde6d31.png

 

Finally, my Shinobi docker settings

Unraid_Shinobi_Settings.thumb.png.1ecd1b12856f152f52d168e165865cf4.png

 

From my Unraid tower, I'm able to ping the Shinobi IP of 192.168.50.10.  However, when I try to launch the GUI, it just spins and finally states that it can't reach the page.  

Prior to trying to set up the managed switch & VLAN, I was able to get to the gui when everything was connected/set-up using the normal network.  I know in the videos SpaceInvaderOne did mention that he was going to have another video showing the VLAN setup but I can't find it.  Also, I'm not currently using PFSense.  I don't know if my current setup (Asus wireless router) supports the VLAN.

 

Any assistance with this would be greatly appreciated.  I'm trying to set this up in a secure setup.

Thanks

Hi, The Asus RT-AC68U doesn't support VLAN's with the stock firmware, you'll need FreshTomato for that! Google will tell you how! Do your cameras support VLAN if not you will need to set the ports they're connected to to tagged and set the PVID. 

Link to comment
6 hours ago, CanuckleHead said:

Thanks for the responses.  I'll look at flashing my wifi routers.  I had done that in the past but reverted back as I tried out the AiMesh feature.
 

I'll also look at the camera (SV3C ProHD 5).  I'm pretty sure I'll have to go down the route of setting the ports as tagged & setting up the PVID.
 

 

or get rid of them and run pfsense off a vm in unraid.

Link to comment
On 9/1/2020 at 4:46 PM, CanuckleHead said:

Hello,
I am in need of assistance trying to get all this working with a VLan.  I'm trying to follow the videos from @SpaceInvaderOne.  Here is my setup:

Wireless Route

 

I bought a tp-link TL-SG108PE managed switch.  This is how I've set it up:

 

Router_VLAN.png.6cebc4c4ac754d89215c11d0a494424f.png

 

 

 

Any assistance with this would be greatly appreciated.  I'm trying to set this up in a secure setup.

Thanks

Hi, 

 

I've never used a Vlan, however I think I can understand how they work. Here are some observations. From that configuration in your Switch you are enabling the VLAN, however in the first table you are setting all the ports as "non members" and in the second one,  you are setting all the ports to ve in the VLAN 1 and you are also setting the VLAN ID to 50 u 1-3,8. As far as I'm concerned ports 1-3,8 should be excluded from VLan 1 and left alone in the VLan 50.

In space invaders video he has his cameras set to a fixed IP with a different topology from the rest of his network. If you configure your router to DHCP it will only assign ip adresses to one topology only, so you should assign each camera one (or set another DHCP server into that VLan 50)

In my case, should it help you, I have 2 separate networks, one is from my home and the other is for the cameras only.

My server (now a Dell T420) has two ethernet adapters, one connects to my home network (192.168.2....) and the other connects to the cameras (10.10.10...) for that matter I enabled docker app called "Glass | ISC DHCP Server" Which listens to my cameras and assigns a dynamic IP address when requested. 

I hope I have at least pointed you towards something.

Link to comment

Now posting from my concerns. I will leave the "hardware acceleration" issue aside I ended up upgrading my crappy E5 2403 to an E5 2470 V2 and so far it runs way better.

 

However, I Still have issues with the playback, and I don't know why I can only see 2 days of videos. The videos are in the hard drive, but the video player does not see them. In the admin settings, they are set to store up to 40 days.

Here are some screenshots you like:

Screen Shot 2020-09-03 at 11.19.38 AM.png


Screen Shot 2020-09-03 at 11.20.20 AM.png
The second issue, as you may have noticed is that one of the "test" cameras only show a green screen in the recording (very useful by the way in a surveillance system). 

So, here are my new 2 questions:

1.- Why can't I see more than 2 days in the video player?

2.- Has the green screen occurred to you ? My cameras are Reolink RLC410 and Reolink and RLC410 5 MP.

Thanks in advance for your valuable input.

Edited by rojarrolla
Link to comment
2 hours ago, rojarrolla said:

Hi, 

 

I've never used a Vlan, however I think I can understand how they work. Here are some observations. From that configuration in your Switch you are enabling the VLAN, however in the first table you are setting all the ports as "non members" and in the second one,  you are setting all the ports to ve in the VLAN 1 and you are also setting the VLAN ID to 50 u 1-3,8. As far as I'm concerned ports 1-3,8 should be excluded from VLan 1 and left alone in the VLan 50.

In space invaders video he has his cameras set to a fixed IP with a different topology from the rest of his network. If you configure your router to DHCP it will only assign ip adresses to one topology only, so you should assign each camera one (or set another DHCP server into that VLan 50)

In my case, should it help you, I have 2 separate networks, one is from my home and the other is for the cameras only.

My server (now a Dell T420) has two ethernet adapters, one connects to my home network (192.168.2....) and the other connects to the cameras (10.10.10...) for that matter I enabled docker app called "Glass | ISC DHCP Server" Which listens to my cameras and assigns a dynamic IP address when requested. 

I hope I have at least pointed you towards something.

Hi, with TP-Link easy smart switches you can’t remove any ports from main vlan, the first table in the screen shot is a bit of a red-herring as nether vlan is selected so the table isn’t showning anything, with vlan 50 only the trunk port and the port connected to server should be untagged the ports to the cameras should be tagged and pvid set to 50. The router with FreshTomato firmware supports vlans and multiple DHCP servers, it’s what I’ve been using for 5 years or so and works well. :)

  • Like 1
Link to comment
39 minutes ago, MothyTim said:

Hi, with TP-Link easy smart switches you can’t remove any ports from main vlan, the first table in the screen shot is a bit of a red-herring as nether vlan is selected so the table isn’t showning anything, with vlan 50 only the trunk port and the port connected to server should be untagged the ports to the cameras should be tagged and pvid set to 50. The router with FreshTomato firmware supports vlans and multiple DHCP servers, it’s what I’ve been using for 5 years or so and works well. :)

Thanks for the information.  I'm learning more than I thought I would need about home networking.  I might try the FreshTomato firmware on my wifi router to see if I can get it working but the more I think about this, the more I think I should improve my home network topology with maybe a dedicated router (Ubiquiti has lots of options).  I might look into PFsense but the routers aren't that costly.  I would then try to use my 2 Asus routers as access points but might swap them for new access points. 

 

Link to comment
8 minutes ago, CanuckleHead said:

Thanks for the information.  I'm learning more than I thought I would need about home networking.  I might try the FreshTomato firmware on my wifi router to see if I can get it working but the more I think about this, the more I think I should improve my home network topology with maybe a dedicated router (Ubiquiti has lots of options).  I might look into PFsense but the routers aren't that costly.  I would then try to use my 2 Asus routers as access points but might swap them for new access points. 

 

I would try the Tomato firmware first, freshtomato is currently still updated and gives you all the features that UniFi or pfsense do and won’t cost you any more! If you want better Wi-fi then ubiquity is hard to beat! :)
 

Link to comment
On 4/14/2020 at 10:14 PM, Oxxy said:

Hi All

 

I am running into performance issues. Running an i7 with 24GB RAM, and 8 Hikvision cameras. CPU is mostly pegged at 100%

For the stream i have it set to copy, with no audio, ditto for recording.

 

I have followed SpaceInvaderOne's guide on passing through an nVidia GPU to try and leverage HW acceleration, ditto with QuickSync - still ends up with 100% usage.

 

I'm writing to an unassigned drive, as per the guide (4TB Purple)

Stream Type - Poseidon, via WebSocket

 

I have tried

Video Codec ...

Copy

H.264 NVENC

H.264 QuickSync Video

H.264 VA-API Intel Acceleration

 

I have tried these settings in 'Input' and 'Recording', but now i feel like I am going around in circles. Is there something silly i am missing?

 

Also, when i used any sort of HW acceleration setting, the Monitor windows are all black.

 

I'd like to move to Unraid an Shinobi, but not if 8 cameras are going to peg an i7 and 100% 24/7.

 

Thoughts?

 

EDIT - just set the entire thing up again on an AMD 1600AF and 16GB RAM, CPU usage isnt 100%, but its way too high to leave like this 24/7

And log for one of the cameras trying to use nVidia card - "Unknown encoder 'h264_nvenc' ", and "Unknown decoder 'h264_qsv'" for Quicksync

 

Hello, I have had the same issue, I even bought an Nvidia 1050 card to help my CPU and I couldn't make the CPU workload go down to a 5 or 7%, it remained with only 2 reolink (RLC410 and RLC410-5MPx) cameras at 2 Mpixels at around 30%.

 

So I upgraded my cpu and I am looking forward to add a second cpu (I have a Dell T420 Server with a E52470 V2 processor which is the top for this kind of server).

 

Now with 2 cameras at 4 and 5 Mpixels, the CPU is around 11-13% with motion detection on and no HW acceleration (because if I enable it the screen goes black and the cameras won't record anything).  Shinobi has been running for 12 or 15 hours only (on this test, I've made lots or test runs) . If that is all the load for each 2 cameras I add, I think I will stick to Shinobi since I need only 5 cameras on this system.

 

However when I started this last test run, the CPU was working between 9-10%. So My concern, and I saw it before with my old CPU is, that, the more time it passes by, the higher the CPU workload gets, and, after around 2 weaks, the CPU workload is so high (around 78% 80%), that the server becomes unstable and almost inoperable.  

 

I don´t know if this happens also to you, and I don't know if you could actually enable the hardware acceleration with motion detection correctly. Did you make it happen?

 

Cheers

 

Link to comment

I want to thank @SpaceInvaderOnefor all his work with this container template. Great job.

I've read the whole thread before posting, so I don't think anyone else has already mentioned this, even though I've seen others referring to similar issues. It seems the container is not saving any changes to the conf.json file. I've tried several times adding my subscriptioId to the conf.json file, only to find out that it was being overwritten with the sample, whenever the container is restarted. I'm new to both Unraid and Docker, so I was unable to fix the problem myself. I've noticed there is a init.sh script somewhere and it does exactly that (copy the sample to the conf.json file), so it seems to me something in the container definition is missing and that makes init.sh if clause to go the wrong way. Sorry if I did not make myself clearer, because English is not my native language.

Link to comment
15 hours ago, cesarvog said:

I want to thank @SpaceInvaderOnefor all his work with this container template. Great job.

I've read the whole thread before posting, so I don't think anyone else has already mentioned this, even though I've seen others referring to similar issues. It seems the container is not saving any changes to the conf.json file. I've tried several times adding my subscriptioId to the conf.json file, only to find out that it was being overwritten with the sample, whenever the container is restarted. I'm new to both Unraid and Docker, so I was unable to fix the problem myself. I've noticed there is a init.sh script somewhere and it does exactly that (copy the sample to the conf.json file), so it seems to me something in the container definition is missing and that makes init.sh if clause to go the wrong way. Sorry if I did not make myself clearer, because English is not my native language.

Copy conf.json to appdata/shinobipro/config/conf.json and Make changes there.  On startup this config overwrites the default one.

Edited by Illuvatar
Link to comment

Does shinobi pro require a tremendous amount of ram?

Quote

Sep 16 22:57:35 Cube kernel: ffmpeg invoked oom-killer: gfp_mask=0x6000c0(GFP_KERNEL), nodemask=(null), order=0, oom_score_adj=0
Sep 16 22:57:35 Cube kernel: ffmpeg cpuset=ec4525f5d684b50fddd3fadf499eb4df7bd4f03201ed0364afc64e47ca229b02 mems_allowed=0-1
Sep 16 22:57:35 Cube kernel: CPU: 2 PID: 20673 Comm: ffmpeg Not tainted 4.19.56-Unraid #1
Sep 16 22:57:35 Cube kernel: Hardware name: HP ProLiant DL380e Gen8, BIOS P73 05/24/2019
Sep 16 22:57:35 Cube kernel: Call Trace:
Sep 16 22:57:35 Cube kernel: dump_stack+0x5d/0x79
Sep 16 22:57:35 Cube kernel: dump_header+0x66/0x280
Sep 16 22:57:35 Cube kernel: oom_kill_process+0x9d/0x21d
Sep 16 22:57:35 Cube kernel: out_of_memory+0x3c8/0x3f9
Sep 16 22:57:35 Cube kernel: mem_cgroup_out_of_memory+0x93/0xc4
Sep 16 22:57:35 Cube kernel: try_charge+0x1bc/0x69e
Sep 16 22:57:35 Cube kernel: mem_cgroup_try_charge+0x112/0x158
Sep 16 22:57:35 Cube kernel: mem_cgroup_try_charge_delay+0x16/0x36
Sep 16 22:57:35 Cube kernel: __handle_mm_fault+0xa94/0x1193
Sep 16 22:57:35 Cube kernel: handle_mm_fault+0x158/0x1a7
Sep 16 22:57:35 Cube kernel: __do_page_fault+0x271/0x40b
Sep 16 22:57:35 Cube kernel: ? page_fault+0x8/0x30
Sep 16 22:57:35 Cube kernel: page_fault+0x1e/0x30
Sep 16 22:57:35 Cube kernel: RIP: 0033:0x15484d44ac6d
Sep 16 22:57:35 Cube kernel: Code: 48 ff c1 eb f3 c3 48 89 f8 48 83 fa 08 72 14 f7 c7 07 00 00 00 74 0c a4 48 ff ca f7 c7 07 00 00 00 75 f4 48 89 d1 48 c1 e9 03 <f3> 48 a5 83 e2 07 74 05 a4 ff ca 75 fb c3 48 89 f8 48 29 f0 48 39
Sep 16 22:57:35 Cube kernel: RSP: 002b:00007ffd1b821cb8 EFLAGS: 00010206
Sep 16 22:57:35 Cube kernel: RAX: 00001547c6870020 RBX: 000000000006da80 RCX: 000000000000cb52
Sep 16 22:57:35 Cube kernel: RDX: 000000000006da70 RSI: 000055ec9a4bc760 RDI: 00001547c6878000
Sep 16 22:57:35 Cube kernel: RBP: 000055ec9a52d8f0 R08: ffffffffffffffff R09: 00001547c6870020
Sep 16 22:57:35 Cube kernel: R10: 0000000000000022 R11: 0000000000000246 R12: 000055ec9a4b4780
Sep 16 22:57:35 Cube kernel: R13: 000055ec9a4b4770 R14: 00001547c6870020 R15: 0000000000005ab6
Sep 16 22:57:35 Cube kernel: Task in /docker/ec4525f5d684b50fddd3fadf499eb4df7bd4f03201ed0364afc64e47ca229b02 killed as a result of limit of /docker/ec4525f5d684b50fddd3fadf499eb4df7bd4f03201ed0364afc64e47ca229b02
Sep 16 22:57:35 Cube kernel: memory: usage 4194308kB, limit 4194304kB, failcnt 1584663
Sep 16 22:57:35 Cube kernel: memory+swap: usage 4194304kB, limit 8388608kB, failcnt 0
Sep 16 22:57:35 Cube kernel: kmem: usage 64964kB, limit 9007199254740988kB, failcnt 0
Sep 16 22:57:35 Cube kernel: Memory cgroup stats for /docker/ec4525f5d684b50fddd3fadf499eb4df7bd4f03201ed0364afc64e47ca229b02: cache:1172KB rss:4127724KB rss_huge:1972224KB shmem:12KB mapped_file:1188KB dirty:0KB writeback:1320KB swap:0KB inactive_anon:48KB active_anon:4128536KB inactive_file:176KB active_file:132KB unevictable:0KB
Sep 16 22:57:35 Cube kernel: Tasks state (memory values in pages):
Sep 16 22:57:35 Cube kernel: [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name
Sep 16 22:57:35 Cube kernel: [ 20256] 0 20256 394 1 36864 0 0 start.sh
Sep 16 22:57:35 Cube kernel: [ 20372] 100 20372 209798 23402 958464 0 0 mysqld
Sep 16 22:57:35 Cube kernel: [ 20593] 0 20593 54193 11169 1105920 0 0 node
Sep 16 22:57:35 Cube kernel: [ 20634] 0 20634 347085 30961 3284992 0 0 node /opt/shino
Sep 16 22:57:35 Cube kernel: [ 20641] 0 20641 55936 11295 1429504 0 0 node /opt/shino
Sep 16 22:57:35 Cube kernel: [ 19334] 0 19334 52303 9328 876544 0 0 node
Sep 16 22:57:35 Cube kernel: [ 19351] 0 19351 574128 363604 3166208 0 0 ffmpeg
Sep 16 22:57:35 Cube kernel: [ 20624] 0 20624 52241 9354 868352 0 0 node
Sep 16 22:57:35 Cube kernel: [ 20673] 0 20673 575436 366141 3223552 0 0 ffmpeg
Sep 16 22:57:35 Cube kernel: [ 48673] 0 48673 52044 9304 913408 0 0 node
Sep 16 22:57:35 Cube kernel: [ 48697] 0 48697 457888 241621 2244608 0 0 ffmpeg
Sep 16 22:57:35 Cube kernel: Memory cgroup out of memory: Kill process 20673 (ffmpeg) score 349 or sacrifice child
Sep 16 22:57:35 Cube kernel: Killed process 20673 (ffmpeg) total-vm:2301744kB, anon-rss:1448468kB, file-rss:16092kB, shmem-rss:4kB
Sep 16 22:57:35 Cube kernel: oom_reaper: reaped process 20673 (ffmpeg), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB

I'm getting a LOT of these in these in the log lately.  I've got the docker set at 4GB, and it looks like it runs hard up against that periodically.

 

Link to comment

trying to figure out how to prove/disprove this, but apparently the following combination causes some issues:

Shinobi Pro

Reolink RLC-520

main (clear) stream at 25fps 2560x1920 8192kB max data rate

monitor set up to record, with same monitor up on screen.

I get tearing/ghosting issues.

2020-09-18T09-05-47.thumb.jpg.063edad2644b508acfa338e5fd7436b0.jpg

 

been fighting with this problem for the past couple days, which is weird since the cameras have been online for almost 2 weeks without issue.

it's not just shinobi that's showing the problem either; the cameras record locally on an SD Card, and those files are corrupted as well.

i'm thinking, the strain of recording the clear stream locally, the sub-stream (fluent) locally (640x480 7fps 160kB), shinobi recording and monitoring the clear stream, is overloading the camera's internal capabilities? I changed shinobi to watch only on one set of monitors (fluent), record on another set (clear), seems to have maybe calmed things down?

Link to comment
On 9/18/2020 at 10:16 PM, sota said:

trying to figure out how to prove/disprove this, but apparently the following combination causes some issues:

Shinobi Pro

Reolink RLC-520

main (clear) stream at 25fps 2560x1920 8192kB max data rate

monitor set up to record, with same monitor up on screen.

I get tearing/ghosting issues.

2020-09-18T09-05-47.thumb.jpg.063edad2644b508acfa338e5fd7436b0.jpg

 

been fighting with this problem for the past couple days, which is weird since the cameras have been online for almost 2 weeks without issue.

it's not just shinobi that's showing the problem either; the cameras record locally on an SD Card, and those files are corrupted as well.

i'm thinking, the strain of recording the clear stream locally, the sub-stream (fluent) locally (640x480 7fps 160kB), shinobi recording and monitoring the clear stream, is overloading the camera's internal capabilities? I changed shinobi to watch only on one set of monitors (fluent), record on another set (clear), seems to have maybe calmed things down?

I had some issues like yours, even I got green screens instead of images, (look for my previous post) I also have reolink cameras.

I don't know if I fixed it, but at least, I don't get green screens anymore, or at least haven't got them and I have turned off my server for two weeks while I was looking for answers at the shinobi forums.

Go to your camera, and in the admin screen look for the advance properties tab, there and look for the option to reset your camera to default settings. I think It will help you with your ghosting.

 

I hope it helps!

 

Cheers!

Link to comment

I confirm that I'm quitting on shinobi.

Thanks Edward Spaceinvaderone for your long and detailed tutorials on how to install and configure Shinobi, I hope it works fine for you. But for me I'm definitely done with it.

I don't want to bore you with details, but Moe, Shinobi's developer, is a real ass! I won't go into details, but he is way to inpolite. 

Long story short, I went and asked him some questions in order to understand better the software and I he did not want to tell me anything. Maybe he is a good developer, but maybe not, otherwise there wouldn't be that many posts regarding his software. But assuming he is indeed a good developer, he is a lousy seller.

So, I encourage you to look for another solution.

 

In my case I am going for a windows VM and I will install Blue iris and give it a try, maybe is not that "flexible" but at least is easier to understand than Shinobi.

Thanks a lot, I've learned some things however, and if I can help you, I would gladly... Not like others.

 

Link to comment
20 minutes ago, rojarrolla said:

I had some issues like yours, even I got green screens instead of images, (look for my previous post) I also have reolink cameras.

I don't know if I fixed it, but at least, I don't get green screens anymore, or at least haven't got them and I have turned off my server for two weeks while I was looking for answers at the shinobi forums.

Go to your camera, and in the admin screen look for the advance properties tab, there and look for the option to reset your camera to default settings. I think It will help you with your ghosting.

 

I hope it helps!

 

Cheers!

Yea i've gone through the default wipe and config reload a couple of times.  It seems to clear it up for a while, sometimes.  For some reason right now, with 4 cameras running, it appears to be pretty stable.  I'm getting occasional smearing on 2 of the cameras, and i'm loathe to dicker with it right now, as I need them to document some shenanigans of a bad neighbor (there's lawyers involved, and probably lawsuits.) So i'm kind of suffering right now with any glitches that might appear, as it's giving me good enough data for now.

 

One issue I just had happen though: my storage drive filled up and I didn't get a notification.  Is there an option someplace to get a notice that the disk is getting full?

Link to comment
On 9/19/2020 at 4:16 AM, sota said:

trying to figure out how to prove/disprove this, but apparently the following combination causes some issues:

Shinobi Pro

Reolink RLC-520

main (clear) stream at 25fps 2560x1920 8192kB max data rate

monitor set up to record, with same monitor up on screen.

I get tearing/ghosting issues.

2020-09-18T09-05-47.thumb.jpg.063edad2644b508acfa338e5fd7436b0.jpg

 

been fighting with this problem for the past couple days, which is weird since the cameras have been online for almost 2 weeks without issue.

it's not just shinobi that's showing the problem either; the cameras record locally on an SD Card, and those files are corrupted as well.

i'm thinking, the strain of recording the clear stream locally, the sub-stream (fluent) locally (640x480 7fps 160kB), shinobi recording and monitoring the clear stream, is overloading the camera's internal capabilities? I changed shinobi to watch only on one set of monitors (fluent), record on another set (clear), seems to have maybe calmed things down?

I have the same camera's and would get this randomly on 2 of the 6. For me I think it was bandwidth related in the end as the data transfer off the box coupled with high network use (6 other heavy users on the LAN) caused them to drop out like your are. Putting them on their own VLAN with a high VLAN priority sorted it sorted it. Well for me. I haven't had the problem anymore but I only have about a week of testing that fix as I switched to Blue Iris. :-)

Link to comment
On 8/15/2020 at 3:14 PM, rojarrolla said:

I agree with you, Shinobi is way difficult to setup. What CPU usage do you get with that setup? And what resolution are you using?

I am using an old gaming box. i7 4790k with 32gb ram. 2x500gb SSD's in extended volume for the boot disk and storage for the "Current, New and alerts" folders. a 4TB mechanical drive for the stored folder and a final 6TB for long term. 

 

Other than some "fun" getting my head around crossing zones for camera triggering to cut down on the moth activation problem at night, it's working without a fault. I haven't tried out the GPU support yet, but the box has a GTX 970 in it that I will get around to messing with when I have time.

 

CPU utilisation wise it floats from 50'ish to 100% but averages arround a solid 70%. That is with the local client running over remote desktop. Closing blueiris and letting it run as a service drops 15-20% off the utilisation figures for me.

 

When connecting via the web interface just to monitor the camera's it sits around 50% and moves up to 95%'ish when recording the hi quality streams with them all activated. 

 

I am recording at Max resolution for each camera, with audio (5 x Reaolink 520's and 1 x SV3C PTZ camera).

 

And the 5 FPS noted on the screen is the FPS for the multi view monitor screen. The camera's are all recording happily when triggered at 15 FPS.

 

I am using pfsense with its own dedicated interface for the cameras and the cameras on their own VLAN to keep traffic problems down. Switch wise they are connected together with a Cisco SG300 series POE switch.

 

working.PNG

InkedCapture_LI.jpg

 

EDIT:

I was being real dumb and left encoding on in BluieIris. Not needed as the cameras already encode. I changed it to direct to disc recording and CPU usage has dropped to 21%

 

 

 

u1.JPG

Edited by Dragosani
Update with processor utilization info.
Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.