Event does not have recordings / No recordings found for the requested time range [Config Support]: #7593
Comments
you should remove this, this is entirely incorrect and not doing anything currently all of this is wrong and the source of your problem, you should be using presets instead of manual args I'd definitely suggest following the docs https://docs.frigate.video/configuration/camera_specific#reolink-cameras and connect to the restream instead of another connection to the camera directly and using correct args instead of random input_args |
|
Hi! thanks a lot for your reply, so , i started from the documentation :
is it right or are you suggesting to switch back to rtsp mode?
this was causing the same issue, this is why i started to mess around with the args, anyway now i've recreated the same starting point so we can exclude any possible issue . ffmpeg:
input_args: preset-http-reolink
output_args:
record: preset-record-generic-audio-copy so far so good, i will update you in few hours. Thanks a lot for your effort!! |
|
ok, update, config is done as per your suggestion, an event of today dated 6.25 PM (now it's 9.45 PM), cannot be displayed (same error), while the image is visible, log is full of errors : any idea? |
|
You should setup the way the docs recommend and if there are errors doing it that way then show those errors. |
|
Make sure to include latest config after doing that. |
|
Hi ! it's what i did, documentation says : the logs are shown in my previous point, any possible solution ? |
No, you're not following the docs. The docs suggest using the restream in the cameras config and you are not doing this. |
|
so you mean to split the detect and record as per this discussion https://github.com/blakeblackshear/frigate/issues/3235#issuecomment-1135876973 so like this : or to follow what is written in the original docs : sorry but i'm kind of lost over here. Thanks :) |
|
Follow the original docs as they are written, only changing the names to match yours. |
|
sorry... but the original docs are mentioning "According to this discussion, the http video streams seem to be the most reliable for Reolink." , for me it is suggested to use the http stream and i should use this approach, can you be more specific on which approach are you suggesting? Thanks! |
|
ok maybe i understood your suggestions : camera 1 is booting after a while, camera 3 doesn't start at all, logs here below : |
|
You need to remove the live height set to 1920 which is incorrect. Also, check go2rtc logs. |
|
Hi, thanks for the check, line removed, after a couple of reboots all the camera are ok, the problem of the missing recordings is still the same. FRIGATE LOGS Details
GO2RTCLOGS also here i cut the lines with same error : Details
the errors are repeated many times, i've tried to copy all the different ones. |
|
enable debug logs |
|
see below : |
|
You need to wait until there is an error in the stream to provide logs, like the |
|
a couple of days are now passed but i cannot see the error "exit status 1", what should i do? Thanks |
|
That would indicate that things are working as expected, no? |
|
actually no, the original problem, which is that the events doesnt have the recording after a couple of hours are passed is still the same, additionally the logs are full of errors and i don't know how to avoid them. |
|
well then we need to see the full copy of frigate logs to understand what the problem is. Previously, the errors were coming from go2rtc which is why the recommendation was to enable debug logs for go2rtc |
|
ok, unfortunately the logs are repeated too many times, and the reply is too long to be posted here on github, i tried to "delete" the duplicated lines (some of the lines are changing only the fps but the error is the same), so what should i look? which part should i paste to provide you a useful information? |
|
You can use pastebin or similar to paste the entire thing |
|
unfortunately i've exceeded the max number of lines also for pastebin...
|
|
all of your cameras are timing out |
|
yes but which could be the reason? configuration of frigate? hardware? settings? |
|
I'd suggest pasting go2rtc logs |
Details
same error is repeated for all the other fps, i paste only the few last since it's too long for paste bin and github : Details
|
|
Those aren't errors, that's just debug data. Multiple pastebins would probably be simplest to be able to see |
|
UPDATE, i figure it out that the issue could be related to the limited space, today i got this message : could be that the space left is always too low for frigate and then it delete all the clips , is it possible to switch off this "space left" functionality? Additional info : What is your opinion on my analysis? |
|
yes, that is very likely to be the issue.
not directly, this is a "last ditch effort" by frigate to ensure it does not entirely fill the drive. You should just config figure to retain a reasonable amount of recordings for the storage that is available or increase the storage that is available for recording |
|
I've already set the retain to 1 day only, the issue seems to calculate incorrectly the free space, I mean there are gigs of free space and frigate keep deleting the clips but the space left is enough to store 1 day of retained clips. Any possibile suggestions? If there is a way "indirectly" to disable this calculation or adjust it (eg saying, if the left space is less than a fixed amount, like 1giga then delete the old recording)? Thanks |
|
The last config you posted has also, you can use fractions when settings the number of days to retain. |
|
what are the camera bandwidths in the storage page |
|
Good catch, I cleared all the clips and modified to 1 day the retention mode. The bandwidth is 3.53 GiB/hr for 7 cameras, but I m not recording continuously only the events where movement is detected, also I don't need the "NVR" functionality, I just want few clips, this means that 20 gigs could be enough for days of retention (it was like this before this new functionality was included, which by the way, I appreciate a lot). |
you ARE recording continuously according to the last config you posted |
|
actually this may be the answer i was looking since weeks... changed to : i'll let it run for a while and confirm if this is the answer to my question. |
|
so you're now keeping 1 day of 24/7 motion recordings, just to make sure that is clear |
|
from my understanding now i will retain only the clips of motion, which will require a lot less space than before, this should fix my issue , it my understanding correct? thanks |
|
depends how often your camera sees motion, it is not uncommon for there to be unexpected motion if things like camera timestamps or trees have not been motion masked |
|
false positive i can fix with the masks/zones and all the fantastic trims frigate has available, so i'm not so worried. Thanks :) |
|
so, a new day, a new issue. the solution suggested before could be working, but frigate continuously check for storage space and keep deleting the clips... this is repeated many times. but with no luck so far, am i missing something? |
|
can you send a screenshot of your storage page (specifically the camera bandwidth part) |
|
expire_interval only applies to the regular recordings cleanup. In this case the issue is that the drive frigate is recording to almost ran out of storage |
|
camera7 is stuck i don't know why, but the bandwidth is the same of camera1 |
|
right, so assuming camera 1 is the same as camera 7 that is 13 GiB/hr if every camera recorded every second then that won't fit in the remaining storage that was there before recordings are cleaned up. Even though you have motion selected as the mode for 24/7, frigate assumes that the entire hour will be recorded so it can ensure that it is safe from filling the storage in the event that there is a lot of motion during that hour |
|
i understand, this is why i tried to force the "expire_interval" , but without any results.... unfortunately i cannot expand the space since i use Home Assistant in a VM and i do snapshot everyday, so i cannot backup 250 gb of space every day. it seems that frigate records everything , if i go to the camera in the UI i can see the recorded videos , while in the event pages there are no events, it's like that the "active_objects" option is ignored by the system |
no, it is not ignoring active_objects. This is what I was trying to explain yesterday and it seems you misunderstood what I was explaining. Your config is telling frigate that you want to record the last 1 day of every second that has had motion detected regardless of if there is an event or object detected |
it is also telling frigate that when an event is created, you want any of those recording segments that has an active object to be kept for 2 days |
|
sorry i saw the following page in the documentation : in the past i was using only the clips, so what should i do to retain only the "clips" with some active object inside? Thanks |
|
You do exactly what that screenshot shows and you use |
|
so, i tried this and i wasn't able to achieve what i wanted , i tried adding : to see if i was able to trick the system but the behaviour is the same, am i doing something wrong? I also tried to leave only one camera to see if my assumptions were correct and the result is yes, the problem is related with the space.
i can navigate in the whole video, it means that frigate is not following the config or i didnt understand how to record only some small clips where the "active_objects" are detected. |
like I've said before, this value does not affect the storage last ditch effort, and there is no way to affect its behavior. Setting this expire interval is just causing more issues for your setup because you are forcing the recordings to exist longer than they should
this is because you have set such a high expire_interval that you are essentially telling frigate not to clean up the old recordings that no longer should be retained by your config. |
|
removed the expire_interval part of the config, same behaviour, i can see the recording of the past hours even without any event. thanks |
|
you need to give it time to clean up, frigate does not run recordings maintenance as soon as it starts |
|
rebooted, cleaned all the events, waited 1 day, same behaviour. i don't know why. Any ideas? |
|
What is your full config |
|
honestly i don't know what i've done, rebooted one more time and now i see that cameras are keeping only the "active movement" events... the following one is the correct config (it has a retain less than your suggestion) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Describe the problem you are having
Hello everyone, first of all, thanks for the amazing work you have done with frigate, i love it and it's working great since a long time in my home assistant home server. Long story short, i keep getting the following error :
when i look back at certain events, usually they are older than few hours but it changes, can be less than 24 hours,it's happening on all cameras, i cannot detect some strange beheviour based on a certain camera.
What i've done to solve the issue so far :
since i'm always looking forward for improvements, please correct any other issue you may find in the configuration (eg hw acc).
Thanks!!!!
Version
SYSTEM 0.12.1-367D724
Frigate config file
Relevant log output
Frigate stats
sorry for the noob here, i was not able to get these statsOperating system
HassOS in proxmox VM
Install method
HassOS Addon
Coral version
USB
Any other information that may be helpful
i have 2 type of cameras :
The text was updated successfully, but these errors were encountered: