Long delay for valid data from Script/Command

[ciela2]

Trying to get data back from simple bash shell script. Got Nan in the graphs for around an hour before it started working after no change on my part.

Put one of the devices in debug and I could see valid data being posted from the script to the debug log, but not being written to the RRD files, they were showing U upon running rrdtool lastupdate.

Has anyone else seen this, does anyone know what is causing it?

In addition one of four rrd files created by the script has apache as owner & group instead cacti.

Something is off with the script/command logic (this script returns a gauge/percentage/floating point number with two digits in case that matters)

Your input would be appreciated, have been pulling my hair out over this.

Thanks

[cigamit]

Ensure that your system time / timezone matches your PHP timezone.

[ciela2]

The system time and the php timezone match.

The Cacti version is 1.2.16 by the way.
It's currently running on a dev server until we can get everything working as it is on our production Cacti server which is on version 0.8.8h

The graphs using the bash script are running intermittently now - three of the four are mostly white space.
Should probably mention that it takes a long time for the poller to run - usually over 200 seconds (this is v1.2.16) - this is for only 6111 graphs.
The 0.8.8h version which has 11832 graphs can run in under 80 seconds.
Both use the cmd poller.

Appreciate your response, please let me know what other information might be helpful in figuring this out.
Thanks

[ciela2]

The system time and the php timezone match.

The Cacti version is 1.2.16 by the way.
It's currently running on a dev server until we can get everything working as it is on our production Cacti server which is on version 0.8.8h

The graphs using the bash script are running intermittently now - three of the four are mostly white space.
Should probably mention that it takes a long time for the poller to run - usually over 200 seconds (this is v1.2.16) - this is for only 6111 graphs.
The 0.8.8h version which has 11832 graphs can run in under 80 seconds.
Both use the cmd poller.

Appreciate your response, please let me know what other information might be helpful in figuring this out.
Thanks

[ciela2]

Just want to mention that this process is much more involved than the old process of adding externally updated rrd files. We have many such scripts running on our Cacti 0.8.8h instance.

Any suggestions greatly appreciated to resolve this weird issue.

Another anomaly is that the rrd file created by apache:apache and the three rrd files created by cacti:cacti do not graph at the same time. If the apache:apache rrd file is updating the other three are not. If the cacti:cacti rrd files are updating the apache:apache rrd file is not (please see attachment):

Screen Shot 2021-09-03 at 11.35.28 AM.png (298.41 KiB) Viewed 1428 times