(SOLVED) May be a bug in 1.2.16

[Badikov]

Hi!

I have upgraded cacti 0.8.8h to 1.2.16. And find some problem.
First is graphic of device in 0.8.8h and second is graphic from 1.2.16
Errors in log file are the same in 1.2.16 and 0.8.8h

POLLER: Poller[0] Maximum runtime of 298 seconds exceeded. Exiting.

And monitoring shows that device is "on" all time.

Where i can find a decision&

[cigamit]

Your poller is overrunning, so we need to figure out why. Can you post the stats line from your log?

[Badikov]

Server ( prod) with 1.2.16 is down now.
So i can not get log files from it.
But i have test server (with the same cacti version , env. and the same database as on prod.
But it doesn`t connect to prod network (security).
I installed spine.
Choose it like main poller ( in cmd.php there is a recommendation to use "spine instead of cmd.php in production due to the built
in scalability limits of cmd.php."

I shell test spine in prod.
And write about that.

P.S. About spine
i have to setup Multi threading in Collection Settings (Data Collector Information)?
Or for current device must setup in "Number of Collection threads" in Management -Devices ?

[Badikov]

Installation of spine and setting in Collection Settings (Data Collector Information) Threads =10 solved my problem!
Now SYSTEM STATS: Time:33.9790 Method:spine Processes:1 Threads:10 Hosts:698