I was just looking back at my timeline and noticed that from 2am-4:50am I am not getting any recording. I have DB compact/repair set to run at 2am so I am thinking that may be the issue. The DB is on a NVE drive so it should be quite fast storage. I am seeing a similar issue with my other BI servers, but they are only for about 10 minutes. I have 30 cameras on that system with 26TB of storage which is way more than the other 2 systems. Is this expected behavior and is there any way to speed it up? It is a big issue to have almost 3 hours of downtime every day.
Recording hole for 3 hours every morning
- Thread starter davidm2232
- Start date
Yep, the downtime is certainly associated with the DB compact/repair. I have mine set to manual and I do it when I am in the system and watching the cameras, not 2am when the perps are out.
Down at the bottom of the BI console - how many clips does it show? It can get unstable after 200,000 clips. It can take a while if you have lots of tiny files.
Down at the bottom of the BI console - how many clips does it show? It can get unstable after 200,000 clips. It can take a while if you have lots of tiny files.
There are just over 409,000 clips. What is the solution to this? I have it set now to do continuous recording with substreams then kick to the main stream when there is motion. Is there a better way? I doubt our system could handle recording every camera in the main stream and we would not have enough storage. Even now, we only have 3 months. Ideally we would like 6 months but need a larger NAS for that.
Yep, that is the problem.
Are you saving jpgs in the Alerts folder? If so each one of those count as a clip and they add up quickly.
Do you have cut and combine into a single file every X hours checked? If not, then every trigger remains its own clip and that adds up quickly.
If you have images in the alert folder and cut & combine unchecked, then I would check the cut&combine box and make it 1 hour and delete the images in the Alerts folder (or move to another folder) and do compact/repair and see if that brings it down to a manageable number or make it stable.
If that doesn't do it and then stop BI service and delete the DB folder in BI directory and then restart BI.
Unfortunately, all of the alert thumbnails in BI are deleted in this process (it would happen whether you deleted the alert images or not) as the thumbnails are simply pointers to the DB and they are lost whenever the DB is rebuilt.
Are you saving jpgs in the Alerts folder? If so each one of those count as a clip and they add up quickly.
Do you have cut and combine into a single file every X hours checked? If not, then every trigger remains its own clip and that adds up quickly.
If you have images in the alert folder and cut & combine unchecked, then I would check the cut&combine box and make it 1 hour and delete the images in the Alerts folder (or move to another folder) and do compact/repair and see if that brings it down to a manageable number or make it stable.
If that doesn't do it and then stop BI service and delete the DB folder in BI directory and then restart BI.
Unfortunately, all of the alert thumbnails in BI are deleted in this process (it would happen whether you deleted the alert images or not) as the thumbnails are simply pointers to the DB and they are lost whenever the DB is rebuilt.
Well, I deleted some older archived clips and set each camera recording tab to cut the video at 8 hours rather than each clip. DB compact/repair went from 3 hours down to 1 hour 15 minutes. That is much more acceptable. Any tips to cut it down further?