icmping

[Korirwn]

Hi,

I've got a router with a icmpping on and a computer behind with a icmpping behind too.
When the router is down, I want that the icmpping on the computer stops.
How can i do that?

Thanks for you help

[Linwood]

If you mean you want the trigger indicating that the computer is not responding to pings to not fire, you want to set the ping trigger on the computer to have a dependency on the ping trigger on the router; then it will not alert until and unless the router stops alerting. Note that you may want the pings to be faster (or at least the alert after less failure) on the router to avoid a race condition between the two. I set most polls for things like SNMP (anything where a failure to respond is a failure that alerts) as dependent on the self-same device's ping, and then all devices pings dependent on the edge device(s) for a location.

[Korirwn]

Thanks.
But I don't want an alert, i want something like this : " if the routeur is down, don't ping the computer"
it's with SNMP?

[prasad_SL]

hi,

just an idea,
can you disable an item via API ?
Try to write an action script, trigger on "router unreachable", and disable "icmpping computer" item via Zabbix API.

I will test and let you know tomorrow.

[Korirwn]

Thanks for you help

I try this " Try to write an action script, trigger on "router unreachable", and disable "icmpping computer" item via Zabbix API."

[Korirwn]

Hi,

I try this but if i disable "icmpping computer", the ping stops and i need it for a ok test.

[prasad_SL]

hi,

As your original post
"
But I don't want an alert, i want something like this : " if the router is down, don't ping the computer"
"

you want to dissable ping.

Ok,
I think you need to enable and dissable icmpping computer on Router ping ok and fail.

If router ping ok -> ping pc
if router ping fail -> stop ping pc
if router ping recovered -> start ping pc

Am I correct ?

[Korirwn]

Hi,

Yes that's right, I want to do do this.

[prasad_SL]

Hi,

did you try to enable the item from action script on "Router ping ok"
on recovery action

you can trigger action scripts on trigger, trigger ack and recovery.

[Korirwn]

with configuration > action ?

[Linwood]

But I don't want an alert, i want something like this : " if the router is down, don't ping the computer"

I can't help but ask - why? You are jumping through some serious hoops to solve what appears to be a non-problem -- if you can suppress the alert, why do you care if the ping keeps going (and is failing)?

[prasad_SL]

Action->recovery operations

[Korirwn]

Hi,
@Linwood because in reality, it's not one computer but many, so i don't want useless requests

[prasad_SL]

Hi K,
Please see the attachment ..

[Linwood]

Hi,
@Linwood because in reality, it's not one computer but many, so i don't want useless requests

I get the concept, I just offer the observation that such pings (unless you are doing something quite unusual) has negligible network load especially since they aren't getting across the (presumably) slower gap to the remote site. What you are doing is conceptual elegant but implementation is going to be somewhat fragile and convoluted. If Zabbix had it all built in -- well maybe. But having to script the stop and the start present opportunities for either to fail, and especially if the restart fails you aren't pinging and might not even know it.

I'll stop trying to sell that idea, just wanted to offer it.