ZoneMinder Forums

segordon wrote:Sure, that's easy.

gist github com segordon 05f12db6dad483fca114

This is a non-crashing copy of zmfilter.pl correct?
Also, please take a look at the other suggestion in parallel (in my post above)

Okay, that eliminates late DB writes. What remains is a) zmfilter db handle is going stale without generating errors and magically recovering b) the sql command is getting mangled That Info command was to check b. You might want to check where you changed it. That Info should not cause a crash unles...

Okay, I have a theory on this: The only reason I can think of this is happening is that this line https://github.com/ZoneMinder/ZoneMinder/blob/master/scripts/zmfilter.pl.in#L638 is returning null till a "filter reload" happens (which is around that 5 min timeline) This could be three thin...

There is something else going on here. As far as I know: 1. If you enable Events = Yes for a user, you cannot prohibit them from rewinding/pausing/skipping events. There is no code in ZM that restricts this that I found. 2. The area you circled in your image is the "progress bar" --> its s...

Was that showing it executed in 0.01s ?

In that case I am currently officially out of ideas

Yes, you can restore. I'm running out of ideas but one more thing to check that ic0n suggested - lets see how long it takes for your event query DB to run: Can you log into mysql DB and execute this SELECT E.Id, E.MonitorId, M.Name as MonitorName, M.DefaultRate, M.DefaultScale, E.Name, E.Cause, E.No...

Okay, lets see if its a caching issue. Make a backup of zmfilter.pl, then replace all occurrences of "prepare_cached" with "prepare" and run again. Does that help? same behavior, everything getting processed in batches. Forced two alarms at 6:34:08 and 6:35:02, neither were proce...

I tried your suggestion, and added a moveRelDown function, but I still get the same error when trying to use the down control in zmview pro It's hard for me to know what is going on as I have no visibility on what that client is doing. But if you have added a moveRelDown, it should at least not rep...

Still going through the process of elimination here. Let's conclude if the problem is with the action of the filter or the process of identifying filters. Let's do this: 1) Disable your alertEmail filter 2) Create an executable script in /tmp called eventcatch.sh with just this: #!/bin/bash echo &qu...

Okay so far: 1. Zmfilter is not crashing 2. It is being invoked every 20s but your filter is executing late So my suspicion then is either alert filter or the sendmail command. Can you please post how you've defined your filter? I'll take a look tomorrow. As a next step lets modify zmfilter to write...

So one thing I noticed in that large log file which seems to span around 7 minutes is zmfilter already exited with a signal code of 14. 08/11/2015 14:27:33.041486 zmdc[1378].INF ['zmfilter.pl ' stopping at 15/08/11 14:27:33] 08/11/2015 14:27:33.090275 zmdc[1378].INF ['zmfilter.pl ' exited, signal 14...

Also, can you please redo this experiment above with Debug logs on and please post all the logs generated during those times

Very interesting. So if I get your situation correctly: a) IF you run zmfilter manually, everything happens immediately (as expected) b) If zmfilter is scheduled to run by ZM, you see it run every 20 seconds, but for whatever reason, it does not get into creating a notification email till 3-4 minute...

I'd like to test this for Android. Looks great. OhmMega please hold on a for a few days - I'm hoping to get access to a package for ZM that does not need one to recompile from source (the server). BTW, this is not a comment at you, but in general: I'd appreciate it if folks who asked to be testers ...

Ok, so I have some thoughts for investigation: a) Your logs show both situations - cases where zmfilter is "sending notification" followed by "creating notification" within seconds of each other, as well as situations where they are spaced a few minutes apart. Now the thing is in...