Thold sends 5 emails for dead host

[imro]

For Every host that goes down thold sends 5 emails, each a minute apart. My cron and poller interval are set to 1.

Currently I am running:
Cacti version 0.8.7g
Thold version 0.4.2
on CentOS 4
PHP Version 5.2.14
mysql 5.1.50

[cigamit]

You mention your cron interval and poller interval are set to 1 minute in Cacti Settings. I assume your Crontab / scheduled task is also set to 1 minute in the OS?

[imro]

here is my crontab

Code: Select all

*/1 * * * * php /var/www/html/plugins/monitor/fast_poller.php > /dev/null 2>&1
*/1 * * * * php /var/www/html/poller.php > /dev/null 2>&1

[cigamit]

Please remove fast poller, that hasn't been distributed with that plugin in > 3 years I believe.

[imro]

Well, you have put it in there, but OK (this is CactiEZ 0.4 if I am not mistaken)

Do you think that is what is causing it?

[cigamit]

Well, you have put it in there, but OK (this is CactiEZ 0.4 if I am not mistaken)

Do you think that is what is causing it?

Most likely. Since it is re-polling the hosts every time (that was made back in the day when cacti only did 5 minute polls, but people wanted to know if it was down much much sooner than that).

[imro]

Well, that was not it.

I have all plugins up to date, including monitor. Therefore the fast_poller.php is not there. So the crontab was not doing anything.

[imro]

Ok, so I was poking around and found few things:

1. As I mentioned this installation is based on an old CactiEZ (turns out 0.2b and maybe even older), therefore most of the data templates are based on 5 minute interval.

2. From cmd.php it looks like hosts are checked for availability only when rrd needs updating. Based on this and 1. this happens only every 5 minutes for most hosts.

3. thold's dead host notification is dependent on status_event_count, which in turn based on 1. and 2. stays the same for 5 polling intervals for most hosts.

How do we fix this?

[sendtoyaser]

I have same problem. I received 5 host down notification and 1 up notification. This is happening after i changed polar cycle and cron to 1 minute in cacti and /etc/crontab.
Do you find any solution

[sendtoyaser]

Howie or any other cacti guru plz suggest any solution.

[sendtoyaser]

Problem resolved when i switch from cmd.php to spine poller