Update... crash occurs more frequently when more than one camera triggers and sends the URL I use for AI. Ken and I are actively looking at this.
5.3.3 - September 30, 2020
- Thread starter fenderman
- Start date
Still not sure what's causing the crashing for me. I told Ken that I rolled back and would wait and see if others reported the same issue with external URLs.
kklee
Pulling my weight
I briefly tested 5.3.3.2, it was crashing and restarting every 20 to 30 seconds. I've sent a bug report to Ken.
kklee
Pulling my weight
I've been doing some more troubleshooting and found that it's somehow related to the Post Jpeg Quality setting. High quality snapshots with 4MP and 8MP cameras seems to be the cause. The snapshots are captured, but causes BI to crash. I have it running as a Service configured to restart on its own, so it recovers fine, but when it restarts it tends to lose contact with the wireless RCA video doorbell for a while, which is not fine.
The higher the Jpeg Snapshot quality, the faster BI crashes. At 100%, BI will crash 2 - 3 times per minute. Dropping the snapshot quality to 33% extends BI's time to crash to several minutes. This was tested with Record snapshot enabled on a single 4MP camera (Loryta IPC-T5442TM-AS). Changing the camera settings to a lower quality feed (low bit rate, Good vs Best, etc) also helps, but that defeats the purpose of using a high MP camera. Seems to work fine with 2MP cameras though, or maybe I didn't wait long enough for a crash.
Still working on the issue with Ken...
Please make sure you report your issue to Support, the more data he has, the quicker he can resolve the issue.
The higher the Jpeg Snapshot quality, the faster BI crashes. At 100%, BI will crash 2 - 3 times per minute. Dropping the snapshot quality to 33% extends BI's time to crash to several minutes. This was tested with Record snapshot enabled on a single 4MP camera (Loryta IPC-T5442TM-AS). Changing the camera settings to a lower quality feed (low bit rate, Good vs Best, etc) also helps, but that defeats the purpose of using a high MP camera. Seems to work fine with 2MP cameras though, or maybe I didn't wait long enough for a crash.
Still working on the issue with Ken...
Please make sure you report your issue to Support, the more data he has, the quicker he can resolve the issue.
All of my cameras are 2 MP. I'm using cloned cameras to create JPGs via URL commands for AI analysis and I'm getting crashes still. I also use motion sensors to send URL triggers as well and have them disabled to troubleshoot. I've rolled back to the previous stable version but continue to get crashes.
Ken and I just spoke. Fix coming in 5.3.3.4.
kklee
Pulling my weight
Got an email from Ken saying he thinks he's fixed it. 5.3.3.4 seems to be working for me now.
So far 3.4 is running good for me. It does sometimes make a low-res image if things get busy which throws the masking off for the AI since rez won't match in that case lol. Only other bug I found is don't double click the timeline. Things will freeze (though the images will keep updating). You'll have to crash the app to get out.
I e-mailed him about.
I e-mailed him about.
kklee
Pulling my weight
Anyone else able to duplicate the crash if you double click an area on the timeline to jump all the cameras back? At first mine does nothing but if I click anything else it all freezes.
Ken says he's not able to duplicate it and I only see on latest version so if you have it too please report.
Also anyone seeing warnings about the DB? “could not obtain exclusive DB access 1/0/0/0” for things like the nightly compact, repair attempts etc.
Ken says he's not able to duplicate it and I only see on latest version so if you have it too please report.
Also anyone seeing warnings about the DB? “could not obtain exclusive DB access 1/0/0/0” for things like the nightly compact, repair attempts etc.
While testing each recent release that provides hi-res JPGs, I'm running into random crashing and higher than normal CPU with each of them. For now, I'm sticking with v5.3.2.8 but will continue to help Ken with testing new releases.
kklee
Pulling my weight
.6 fixes my two issues. The db lock was related to the timeline bug. I did have a crash this AM on .5 so we'll see if .6 helped with that.
Ken has been great by email! For any wondering it's def worth the $30 a year considering how often he does releases and how much he listens to us.
He even added a feature I asked for right away. I asked if he could make canceled alerts remove from timeline (but stay in alert history to check for FP's). Boom done, .5 added that fwiw so if you cancel an alert it will remove the mark from the timeline. This frees you up to use Flags for other things now that you want to flag.
Ken has been great by email! For any wondering it's def worth the $30 a year considering how often he does releases and how much he listens to us.
He even added a feature I asked for right away. I asked if he could make canceled alerts remove from timeline (but stay in alert history to check for FP's). Boom done, .5 added that fwiw so if you cancel an alert it will remove the mark from the timeline. This frees you up to use Flags for other things now that you want to flag.
kklee
Pulling my weight
Updated to .6 yesterday, still getting an occasional random crash, not sure about the cause since it's seemingly random.
I'm crashing once a day at exactly 8:00am eastern.
Anyone else only having one crash a day? I don't get the time. My db maint runs at 2am, the machine does nothing else at 8am and my clips are only about 40 min long due to size.
Ken seems lost so make sure to keep reporting them and anything you can figure out. Mine doesn't seem tied to motion or anything it's just this one darn time a day lol.
Anyone else only having one crash a day? I don't get the time. My db maint runs at 2am, the machine does nothing else at 8am and my clips are only about 40 min long due to size.
Ken seems lost so make sure to keep reporting them and anything you can figure out. Mine doesn't seem tied to motion or anything it's just this one darn time a day lol.
