Thold 2.x - Thresholding/Alerting module for cacti 8.6

[twelzy]

torstentfk,

I want to know if your modifications to detect reeboots works with cactid!

I've just changed all the files (following your directions), but the uptime and reboot fields for all hosts are 0 in the database...

Thanks in advance!
(...and sorry for my poor English!)

[twelzy]

Anyone?!

torstentfk,

I want to know if your modifications to detect reeboots works with cactid!

I've just changed all the files (following your directions), but the uptime and reboot fields for all hosts are 0 in the database...

Thanks in advance!
(...and sorry for my poor English!)

[torstentfk]

Hi,

just returned from holidays...
the boot-moni only works with cmd.php, because there the device is checked and stored into the sql-db (simple php-statement); in cactid I must call c-function which are unknown to me.
reboot: hosts=0; this is normal; only if a host is not reachable, there is an one stored for only ONE cacti-polling cycle. Thold notices this and emails. Then the flag (not counter) is reseted.

Torsten

torstentfk,

I want to know if your modifications to detect reeboots works with cactid!

I've just changed all the files (following your directions), but the uptime and reboot fields for all hosts are 0 in the database...

Thanks in advance!
(...and sorry for my poor English!)

[klr0514]

When I attempt to autocreate thresholds, I get the following error:

Update failed:
INSERT INTO thold (element, rra) VALUES (, 76)
You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near ' 76)' at line 1

I am running with the latest thold and the plugin architecture. I added the database structure (not sure if this should have been added under cactidb or not) using thold.sql, but still nothing. The thresholds tab is working, and all devices are showing up, but that's all I get.

[klr0514]

Interesting... Thresholds are working for my localhost (hard drive space, memory, etc), but not working for the network devices I am monitoring (routers, switches)... Any ideas?

[kanada]

Hi,

just returned from holidays...
the boot-moni only works with cmd.php, because there the device is checked and stored into the sql-db (simple php-statement); in cactid I must call c-function which are unknown to me.
reboot: hosts=0; this is normal; only if a host is not reachable, there is an one stored for only ONE cacti-polling cycle. Thold notices this and emails. Then the flag (not counter) is reseted.

Torsten

Can You msg us, when You find solution? i test that on my cacti and You have right. But...

I dont know what now, couse i have about 200 monitoring hosts, on cmd system check it about 3 min on cmd.php, and 30 sec on cactid. I cant monitor any host using Thold, couse its takes too long.

[torstentfk]

Hi,

just returned from holidays...
the boot-moni only works with cmd.php, because there the device is checked and stored into the sql-db (simple php-statement); in cactid I must call c-function which are unknown to me.
reboot: hosts=0; this is normal; only if a host is not reachable, there is an one stored for only ONE cacti-polling cycle. Thold notices this and emails. Then the flag (not counter) is reseted.

Can You msg us, when You find solution? i test that on my cacti and You have right. But...
I dont know what now, couse i have about 200 monitoring hosts, on cmd system check it about 3 min on cmd.php, and 30 sec on cactid. I cant monitor any host using Thold, couse its takes too long.

Hi,
we have 1400 tholds and it runs very quickly. cmd polls 2300 rra files.
>
>04/19/2006 11:35:38 AM - SYSTEM STATS: Time:37.4910
>Method:cmd.php Processes:3 Threads:N/A Hosts:264
>HostsPerProcess:88 DataSources:3818 RRDsProcessed:2307
I have problmes with cactid that cactid get not all data from the polled devices in unregulary intervalls. So I must use cmd.php and it runs very fast; also I don't have enough knowledge for programm C with mysql support. So someone else must take this part (developer?)


Torsten

[kanada]

Hi,

just returned from holidays...
the boot-moni only works with cmd.php, because there the device is checked and stored into the sql-db (simple php-statement); in cactid I must call c-function which are unknown to me.
reboot: hosts=0; this is normal; only if a host is not reachable, there is an one stored for only ONE cacti-polling cycle. Thold notices this and emails. Then the flag (not counter) is reseted.

Can You msg us, when You find solution? i test that on my cacti and You have right. But...
I dont know what now, couse i have about 200 monitoring hosts, on cmd system check it about 3 min on cmd.php, and 30 sec on cactid. I cant monitor any host using Thold, couse its takes too long.

Hi,
we have 1400 tholds and it runs very quickly. cmd polls 2300 rra files.
>
>04/19/2006 11:35:38 AM - SYSTEM STATS: Time:37.4910
>Method:cmd.php Processes:3 Threads:N/A Hosts:264
>HostsPerProcess:88 DataSources:3818 RRDsProcessed:2307
I have problmes with cactid that cactid get not all data from the polled devices in unregulary intervalls. So I must use cmd.php and it runs very fast; also I don't have enough knowledge for programm C with mysql support. So someone else must take this part (developer?)


Torsten

Yes, on my little box simply request on cmd.php take about 60 sec.
That same on cactid take about 3 sec.
When host is down, system wait about 2-3 cycle and then send notify to mail.. i get message about 15 min after downing. Its take to long..
I create new cacti on other host.. i set only one host to thold, i get notify 2-3min after host down.

Any ideas?

[chadd]

Does anyone know why adding a cdef (make per 3 min) would cause my tholds to stop triggering? The graphs are still showing data that is in exception to my thold triggers, but they are not going into alarm. I know they trigger off of the data sources, but as long as there is valid data in the rrd, why would this change make tholds stop working? Anyone?

Thanks,

-chadd.

[cigamit]

Does anyone know why adding a cdef (make per 3 min) would cause my tholds to stop triggering? The graphs are still showing data that is in exception to my thold triggers, but they are not going into alarm. I know they trigger off of the data sources, but as long as there is valid data in the rrd, why would this change make tholds stop working? Anyone?

Thanks,

-chadd.

How is the CDEF changing the graph data? Can you show us the equation that it does? You say your graph is showing that it should alert, but with a CDEF what the graph shows and the true value are different (depending on the CDEF).

My current SVN version actually let you apply a CDEF against the thold value before returning it (as long as its a simple single datasource CDEF, also no SIN/LOG/etc... functions yet) which makes it easier when you are dealing with graphs and CDEFs and want to report on the same scale of info.

[chadd]

cdef=CURRENT_DATA_SOURCE,180,*

My poller runs every 3min.
Just so you know, the graph template that I am playing with is the interface error/discards template.

I also changed the 'GPRINT type' to exact numbers - from 'NORMAL'. What I was trying to accomplish is to get the actual number of errors per/poll rather than a number like '8.2m or (8.2 * 10^ -3)'. What I would like, is just, '8 errors that 3min'.

LVM suggested the cdef in another thred, which is why I am doing that. It seems to be working... kind of.. But I am still not getting an email every time there are errors on the graph (thold is set to 1 for all interface error/discard datasources).

Thanks for any help you could give. I will try your latest version as well. Are there any corrections I will have to make in your latest version - or is it pretty much remove the old php and install the new? Thanks again.

-chadd.

[cigamit]

You should first try running thold from the command line to see what it gives you.

(while in the thold directory)

Code: Select all

php -q check-thold.php /debug /show

[venquessa]

hostname cacti # php index.php
Segmentation fault

What is it with this segfault carry on? It's the only PHP program I have ever known to ignorantly segfault with no warning nor error.

I've spent an age going through the install with a fine toothed ..

print "Here we are!";

trace

Until I found the segfault in the db_fetch_row function, called from the read_config_something function.

No reason could explain why it segfaulted in the cacti_log calls.... or appeared to, but only on some calls not others. So I unwound the add on with the uninstaller and got my cacti back + a huge hole in all the graphs.

I realise it's probably a simple versioning problem with some function of other and not your fault exactly, I can patch your add-on in by hand I believe. Might even be a PHP/MySQL/Apache/Cacti version problem on my end.

"Good evening sir. Would you like errors with that SegFault?"
"Oh yes please!"

My setup:
Gentoo 2005.2
apache-2.0.55-r1
PHP 4.4.2-pl0-gentoo
cacti-0.8.6h_p20060108-r2
MySQL 4.0.24

[cigamit]

hostname cacti # php index.php
Segmentation fault
<snip>
No reason could explain why it segfaulted in the cacti_log calls.... or appeared to, but only on some calls not others. So I unwound the add on with the uninstaller and got my cacti back + a huge hole in all the graphs.

Seems like you are running the very old version of thold, since it has an "uninstaller". I would recommend running the Plugin Architecture, and adding the newest version of the thold plugin (links in sig). The old version hasn't been updated in ages.

[ServerTweak]

hello

what lines to add the cacti path in the install.sh?