-
agent_keepalive alerts not happening
-
Hey Nil,
2 – I have an alert setup to check to see if the agent is down. The min/max value is set to 1. if anything other then 1 send an alert. With a Min Alert of 1 and a Max Alert of 5. Time threshold of 5 min. It was working fine until I restarted the agent in question and the alert’s stopped however I did not get the [RECOVERED] message that is supposed to be delivered.
I take the agent down today and the alerts do not fire.
Any idea what could be the problem?
-Al
BTW Well done on a great product!You have any Network Module (of any type) that are currently updating the agent ?, be sure also that you have a keepalive module assigned to this agent. Could you post a screenshot ? of agent main page ?
Thanks.
-
-
-
Sorry for the lateness in my reply. I put a bug in on soureforge with the screen shots. As for my version I am using the latest from the SVN downloaded the branches and the trunk. So if it’s been updated there I have the latest and greatest.
Good idea to put the screenshots in bugtracker.
I have seem two things very interesting.
First one:
You’ve time threshold set to 5 min, and a min of alerts of 1. Probably your agent interval is set to 5 min also. This means that when agent goes down, at 5 mins (t1) a internal alert is fired but not visible for system because there is a minimun of 1 alert, and when the second one comes, time threshold is above the limit, so counters are reset: this alert will never fired up.
Set time threshold to 30 and try again. This should work. Combination of factors in alerts are always confusing 🙁
Second one and most easy to understand.
You have defined a text alert ! :-). Check out the third line on alert definition “alert text” is for parsing a text data with regexp, not to parse numerical data. This makes alert to search for strings not for compare integer. This is for sure the problem.
Hope this helps !
-
Hi nil,
alerts are fired if I put min/max values to 0/0 in the alert configuration. But This was just for a test. the agent_alive never can fire an alert at my system, because the module itself never changes the value from 1.
If I stop the pandora – service on a server, the agent_keepalive module stay at value 1 and doesn’t turn to 0.
No matter how you define your alert, the module will never fire anything.so how do I have to configure the agent_keepalive module that changes between 1 and 0 if the service is stopped or running?
-
-
Nil,
I think the problem might be a little more fundamental.I dropped the database and I am recreating it and I am finding a number of bugs associated with the pandoradb_data.sql
I will list them here:
First line 67 Syntax error:
Was:
INSERT INTO `tlanguage` VALUES (‘fr’,’Français’)
changed to:
INSERT INTO `tlanguage` VALUES (‘fr’,’Français’);
Second line 207
ERROR 1062 (23000) at line 207: Duplicate entry ‘100’ for key 1Commented out line 207
I am not sure if any of the above could be my problem because I didn’t see the error’s the first time I setup the database but I am finding it this time around.
Also it seems like reloading the database cleaned up my icon problems.
Interesting…
-
Hi!
I took a look at the problem you had.
Yes, there was a “bug” in the file pandoradb_data.sql file in trunk and branches/Pandora_1.3.x branches at the SVN:
There was not a “;” at the end of line 66 and 67, respectively. Also, last line in branches/Pandora_1.3.x had a value that was inserted at line 165. This only gives some warning because that line was added before, no problem at all, but a bit annoying.
Thanks for reporting this. I have updated the SVN.
Raúl
-
-
-
Alerts are still not working though.
Starting to get very disappointed…
🙄Let’s begin again.
a. Make sure that Keepalive module is enabled in this agent.
b. Make sure that agent DO NOT HAVE any network module
c. Make sure that alert defined for this agent is set max 1, min 1 for keepalive.
d. Make sure that that agent is going from ALIVE (have data coming from AGENT inside the interval)
e. Make sure that alert threshold is at least twice times bigger than agent interval.
f. Make sure that agent min. alerts is 0 (could have other values, but with 0 it’s more easy/quick to test).
g. Make sure that agent max. alerts is 1 (by the same reason that f.) -
-
-
Ahh nice, it’s working.
My network modules were the problem. Thx a lot nil for the step by step manual … these kinds of little mistakes … i hate em.
I just deleted my network modules and pandora fires the alerts of the keep_alive module.Pandora really rocks !
Thanks thanks :-). I hope this problem don’t happen again thanks to your patience, and because now it’s well documented !
-
-
Ok it’s working now!!!
Apparently it was the Min Agent that was causing the problem. On the wiki for configuring an alert it said to have Min Alert set as 1 and max alert set as 1
So it looks like some documentation issue….
-
Ok I know I am not retarted but anything is possible.
Downloaded the VMWare image and loaded and got pandora started.
However it only fires 1 alert and never repeats the alert.
I have this on my test agent
And I have seen this on the ROUTER agent that is configured with this release of Pandora 1.3.1
Even on the router it isn’t firing more then 1 alert.
Suggestions please?
thanks,
-Al
