Hell Yeah! Direct Deepstack Integration - 5.4.0 - March 31, 2021

Found the performance tab in task manager. Regardless of the 100% heading, looks like I can have 100Mbit/s on each ethernet port?. Just updated BI to latest and notice the memory shot up to nearly 100% and now PC has crashed. Restarted, same problem.
Got rid of 5.4.5.0
 
@Dave Lonsdale - yeah on that machine and 1 image sent to DS, I would expect it to not max out.

I suspect a setting is wrong on your network cards to limit it to 100MB - they are probably 1gig cards that got reset after an update or something crazy.

I had that happen once or twice. Go into the network settings and many times the speed of the card is right in the name of the card you have. Hopefully it says giga.

If so, disable the card momentarily and then turn it back on. If that doesn't fix it, then do the Diagnose button and see if that corrects it.

If you indeed have a 100MB card and you are seeing performance issues, you can pick up a giga card fairly cheap and replace it.
 
Thanks again for the feedback wittaj. I didn’t find a speed selection for my Intel Gigabit 4P i350-t adapter card and I imagine it would automatically operate at 1 Gigabit to match its partner. But my simple switches are 100Mbit/s in and out and I now see that one of the ports, although varies, is running at an an average of approx 88Mbit. Me-thinks I need to dip into thee piggy bank again for a third switch.

BI has already been updated and runs OK in the PC with memory use as previously.
 
  • Like
Reactions: sebastiantombs
Pretty much once a day it starts showing and leaving some of the snapshots in the clip menu on the left of BI. I have tried using 'alerts' and 'confirmed alerts' filters with the same result. Any ideas?
 
Thanks again for the feedback wittaj. I didn’t find a speed selection for my Intel Gigabit 4P i350-t adapter card and I imagine it would automatically operate at 1 Gigabit to match its partner. But my simple switches are 100Mbit/s in and out and I now see that one of the ports, although varies, is running at an an average of approx 88Mbit. Me-thinks I need to dip into thee piggy bank again for a third switch.

BI has already been updated and runs OK in the PC with memory use as previously.
Even if they are giga, your are probably running in auto and it may be auto-negotiating to the 100mbs of the switch. At first I thought you meant your were using link aggravation was going to suggest you double check your aggregation settings. you might be able to set there speed under the driver/device properties in device manager.
 
Even if they are giga, your are probably running in auto and it may be auto-negotiating to the 100mbs of the switch. At first I thought you meant your were using link aggravation was going to suggest you double check your aggregation settings. you might be able to set there speed under the driver/device properties in device manager.
Sounds like his switches were 100Mbps only. You should almost never need to manually set a link speed anymore. If he gets a gigabit switch, his bandwidth problems will magically disappear without any intervention.
 
Sounds like his switches were 100Mbps only. You should almost never need to manually set a link speed anymore. If he gets a gigabit switch, his bandwidth problems will magically disappear without any intervention.
Thanks for the feedback guys. My network knowledge is extremely shallow to say the least but I imagine that solving my overloaded ethernet port by replacing my 100Mbit 16 port PoE+ switch with another one that has a 1 gigabit uplink would be far more expensive than offloading some of the cameras on to another smaller 100Mbit switch and then simply using another port in the 4P i350-t.
 
Thanks for the feedback guys. My network knowledge is extremely shallow to say the least but I imagine that solving my overloaded ethernet port by replacing my 100Mbit 16 port PoE+ switch with another one that has a 1 gigabit uplink would be far more expensive than offloading some of the cameras on to another smaller 100Mbit switch and then simply using another port in the 4P i350-t.
Your confusing me with the offloading. I don't fully understand your hardware configuration maybe draw it out. So I was re-reading your post on the previous page so your using three NICs, built-in nic for internet, and 2 of your additional 4 ports connected to two 16 port poe switches then your cameras? How many cameras are your talking about? Do you need 16 ports?
 
Last edited:
As an Amazon Associate IPCamTalk earns from qualifying purchases.
With last update 5.4.5.1 have continued the missing alerts.I see the log and gives DeepStack: Alert cancelled (nothing found) . I have test 20 images for analysing but have the same problem. For this reason i am runnig paralel the AI tool to compare the alerts.I thing is BI issue.
 
With last update 5.4.5.1 have continued the missing alerts.I see the log and gives DeepStack: Alert cancelled (nothing found) . I have test 20 images for analysing but have the same problem. For this reason i am runnig paralel the AI tool to compare the alerts.I thing is BI issue.
When you right click snapshots and do Analyze with DeepStack, do you still get nothing?
 
When analyzing with right clicking getting Normal answer as person detected.
At 20 triggers 5 going DeepStack: Alert cancelled (nothing found) . Not all.
 
Last edited:
When analyzing with right clicking getting Normal answer as person detected.
At 20 triggers 5 going DeepStack: Alert cancelled (nothing found) . Not all.
Then it’s probably your per-camera Trigger > AI settings. Can you post a picture of that AI window for this camera? Kind of hard to understand. So out of 20, you’re getting 15 successes and 5 not found?
 
  • Like
Reactions: maximosm
Does this still happen if you drop the confidence to 1%?

You’re getting 15 out of 20 to work properly? Or none of them?
 
Does this still happen if you drop the confidence to 1%?

You’re getting 15 out of 20 to work properly? Or none of them?
Yes i have also tried this. And geting 15 out 20. At start was thought that was the pre trigger buffer that i had 7sec and thinking that the BI starting to count images for analyse at the moment that starting the pre trigger. I set the buffer to 2 sec but having same problem.
 
Yes i have also tried this. And geting 15 out 20. At start was thought that was the pre trigger buffer that i had 7sec and thinking that the BI starting to count images for analyse at the moment that starting the pre trigger. I set the buffer to 2 sec but having same problem.
Why are you processing so many? Is it possible your overtasking?
 
Why are you processing so many? Is it possible your overtasking?
For this 2 cameras use as trigger from onvif of cameras with a door contact and pir conected to cameras.From the time that door open or trigger from pir take 5-6 sec to see person at the camera.So i play with the number of images for proccesing.Normal must have 6-7sec but i am try to find the problem with this #Alert cancelled (nothing found) # and playing with more images .
 
For this 2 cameras use as trigger from onvif of cameras with a door contact and pir conected to cameras.From the time that door open or trigger from pir take 5-6 sec to see person at the camera.So i play with the number of images for proccesing.Normal must have 6-7sec but i am try to find the problem with this #Alert cancelled (nothing found) # and playing with more images .
ok but I thought BI+Deepstack only worked with BI motion triggers and ignores ONVIF triggers. I am still fine tuning my BI triggers. Usually my CAMS see the motion first send onvif trigger but BI does nothing when BI see the motion then I see the deepastack check. Anyways I am wondering if your 5 misses are because they are onvif triggers and not bi.