Cameras Go Offline Randomly, Changing On Each Restart

looney2ns

IPCT Contributor
Joined
Sep 25, 2016
Messages
15,521
Reaction score
22,657
Location
Evansville, In. USA
I'm having somewhat similar issues with my Blue Iris 5.x install (5.0.1.0 to be exact). It runs as a service but I do sometimes connect via RDP and run the app as well to monitor the feeds. I don't normally use the web interface.

From time to time, maybe after 24-30 hours of uptime, about half of the cameras go into "no signal" and the only solution is to close the console and stop the service, then start the service back up. About half the time the service can't be stopped at all and I have to reboot the machine.

My cameras are a hodge podge of Amcrest, Dahua and Reolink (I read through this thread and was surprised to hear of issues with Reolink "tearing" since I've never experienced it, but maybe my settings are different, unless I don't understand what that means? I think they're using RTSP but I don't remember). The cameras that show "no signal" vary, not any specific make or model.

This had all been working great up until recently, and the recent change I made was that the Intel Driver & Support Assistant I have installed on there prompted to update the Intel HD 630 drivers. I'm now running version 26.20.100.7000.

When that happened I thought maybe it was a good time to also update the drivers for the Quadro P2000 in there, which is doing a lot of the HWDEC work. So I updated it as well to Nvidia's version 26.21.14.3170

My guess is that one or the other of those drivers may be the problem so I'll probably roll both of them back and see if a) the problem goes away and b) introduce one back at a time and see which one causes it. Because I'm curious to know.

It does sound like some Intel video driver versions have problems, so I'm headed to the wiki now to see if that version I have is mentioned as good/bad. If that is the problem, then the lesson I learned is, don't update just because, or let someone else do it first and report back.
If it's not broke, don't fix it. ;) Most likely, you would be better off removing the Nvidia completely.
 

WSCam

Getting the hang of it
Joined
Sep 24, 2018
Messages
32
Reaction score
29
Location
Tacoma, WA
If it's not broke, don't fix it. ;) Most likely, you would be better off removing the Nvidia completely.
Lesson learned there. I've already begun taking a "wait and see" approach when it comes to updating Blue Iris itself, generally giving it a few weeks and checking here for any feedback. I should've done the same for the GPU drivers.

I added the Quadro P2000 because the system was struggling with all of the cameras. I have 20 cameras on the system, with an i7-7700K, 16GB RAM, and a bunch of drives (SSD arrays and WD Purple arrays for longer term storage). The P2000 actually does a good job keeping up, with the "Video Decode" hovering around the 50-55% usage mark. It also does some minimal encoding for cases where I can't do direct-to-disk (time lapse), but it's barely noticeable.

Unfortunately without the P2000, I think the i7-7700K was struggling at around 80-85% on average, and if I wanted to play something back it would hit 100% easily. Since adding the P2000, the CPU is normally around 20-25%.

Why a P2000? Well, I had it already and it wasn't being used. Cheaper in the short term than buying a new motherboard, CPU, and memory. All of which I'll do eventually.

UPDATE: I rolled back the Intel HD 630 drivers earlier and just noticed that it's already "no signal" on over half the cameras, so that probably wasn't the issue. Now I'll rollback the Nvidia and see if that changes anything.

If not I'm back to square one and may just reinstall Windows 10 and start over. I did that about 5 months ago when I finally updated to Blue Iris 5 and thought that was a good time to start fresh, but just the past week started having this problem.
 

WSCam

Getting the hang of it
Joined
Sep 24, 2018
Messages
32
Reaction score
29
Location
Tacoma, WA
UPDATE: I rolled back the Intel HD 630 drivers earlier and just noticed that it's already "no signal" on over half the cameras, so that probably wasn't the issue. Now I'll rollback the Nvidia and see if that changes anything.

If not I'm back to square one and may just reinstall Windows 10 and start over. I did that about 5 months ago when I finally updated to Blue Iris 5 and thought that was a good time to start fresh, but just the past week started having this problem.
After rolling back the Intel and Nvidia drivers to the version I had previously, I still experienced the "no signal" issue at various points throughout the day. I remembered one other change I made before this all happened, and that was related to my timelapse cameras. I have 4 and they're clones of the normal cameras but set to trigger periodically instead of motion. The change I made was to set all 4 of them to enable "Limit decoding unless required". I have those cameras in their own group and don't normally show them in the console, so that seemed like a good idea.

Although it seemed to work fine (I honestly didn't notice any difference in CPU/GPU usage), apparently that was the cause of the error. I have no idea why. Those cameras were still in their own group, not showing in the console at all. But I changed them all back, unticking that option, and I've been up for over 48 hours without issue.

I guess if anyone else is having this issue, maybe check if you have any cameras with that option enabled and try unchecking it for a while to see if the problem goes away. If that's a 'cure' for others, then I suppose it's a bug and it can be looked at. I'm hesitant to call it a bug because maybe this is just some weird issue with my system where that feature is doing something unusual. Something with the Nvidia acceleration I have, the fact that these are timelapse/periodic capture cameras that are re-encoding (not direct to disk), and so on.
 

fenderman

Staff member
Joined
Mar 9, 2014
Messages
36,897
Reaction score
21,250
After rolling back the Intel and Nvidia drivers to the version I had previously, I still experienced the "no signal" issue at various points throughout the day. I remembered one other change I made before this all happened, and that was related to my timelapse cameras. I have 4 and they're clones of the normal cameras but set to trigger periodically instead of motion. The change I made was to set all 4 of them to enable "Limit decoding unless required". I have those cameras in their own group and don't normally show them in the console, so that seemed like a good idea.

Although it seemed to work fine (I honestly didn't notice any difference in CPU/GPU usage), apparently that was the cause of the error. I have no idea why. Those cameras were still in their own group, not showing in the console at all. But I changed them all back, unticking that option, and I've been up for over 48 hours without issue.

I guess if anyone else is having this issue, maybe check if you have any cameras with that option enabled and try unchecking it for a while to see if the problem goes away. If that's a 'cure' for others, then I suppose it's a bug and it can be looked at. I'm hesitant to call it a bug because maybe this is just some weird issue with my system where that feature is doing something unusual. Something with the Nvidia acceleration I have, the fact that these are timelapse/periodic capture cameras that are re-encoding (not direct to disk), and so on.
When you clone a camera, blue iris only decodes one stream. This is why you didnt notice a performance gain when limiting decoding on the clones.
 
Top