Simple probe question

RouterOS The Dude
1

Where do I do to set how many good probe tests it takes to change state from down to up?

2

On the polling tab, check the settings of each area, one for global, one for each map and one for each device…
polling.png

3

So, is probe down count value used for probe up count also?

That does not seem to be the case because up status sometimes takes minutes when system was only down 15 seconds or so (my settings are 10s, 5s, 1) and running v3.6 (v4 did not work well for me on W7 64bit)

EDIT: here’s an example, the lines came up 15 seconds later but up status did not until 5 minutes later:
DownToUp.png

4

…my problem was due to negative cache:

http://wiki.mikrotik.com/wiki/Getting_started_with_Functions_and_probes

5

Yeah, finally someone with the same problem as me :slight_smile: BTW I wrote that…

6

Yeah, thanks for that! Franky, if not for people like you Dude’s idiosyncrasies would make me give up a long time ago :slight_smile:

7

Hi guys

I got the same problem too, and would like to get some help about this. :blush:
Since I have lots of router to monitor, each router have to monitor different interface.
I create several probe to match each interface, each probe use a matching function.
ex.
function

<?xml version="1.0" ?> 57 557421 IntChk_Serial0 oid_raw(concatenate("1.3.6.1.2.1.2.2.1.8.",array_element(oid_column("1.3.6.1.2.1.2.2.1.1"),array_find(oid_column("1.3.6.1.2.1.2.2.1.2"),"Serial0"))))

probe

<?xml version="1.0" ?> 13 557437 Serial0 8 if(IntChk_Serial0()=1, "", "down")

So, seem negative cache time cause a probe down for about 5 mins down time.
I tried to modify my founction below

<?xml version="1.0" ?> 57 557421 IntChk_Serial0 oid_raw(concatenate("1.3.6.1.2.1.2.2.1.8.",array_element(oid_column("1.3.6.1.2.1.2.2.1.1"),array_find(oid_column("1.3.6.1.2.1.2.2.1.2"),"Serial0"))),3,3)

But it seems not working, probe still show down status for nearly 5 mins @@"
Can any one give some ideas about this? :blush:
Thanks a lot~!!

8

Your function is snmpwalking 2 times and snmpgetting 1 time. Am I right?
And you change default “negative cache” only for snmpget oid_raw.
You should try to change “negative cache” also for both oid_column funtions, and check wats up.

As I can see, you had modified mem_size funtion which comming with defauld The Dude distribution.

I found that mem_size and your functions are not optimal.
This code

oid_raw(concatenate("1.3.6.1.2.1.2.2.1.8.",array_element(oid_column("1.3.6.1.2.1.2.2.1.1"),array_find(oid_column("1.3.6.1.2.1.2.2.1.2"),"Serial0"))))

can be replaced with more clear (at least I think it’s more clear)

array_element(oid_column_raw("1.3.6.1.2.1.2.2.1.8"),array_find(oid_column("1.3.6.1.2.1.2.2.1.2"),"Serial0"))

and you should also change the code for your need of “negative cache”

array_element(oid_column_raw("1.3.6.1.2.1.2.2.1.8",3,3),array_find(oid_column("1.3.6.1.2.1.2.2.1.2",3,3),"Serial0"))

As for me, I did never use “negative cache” or “positive cache” because of no need. :smiley:

9

Yeap, that’s what I do for try to make sure which interface I am looking for. :smiley:

well…I wrote this function myself without reference any function in dude..
seems I think too much on how to get interface index and use it to get Interface Oper Sataus @@"

Is that means I can set the cache value to 0?
basically, I don’t want the cache too :stuck_out_tongue:

Really thanks for your help~!!

10

No, I mean defaults are ok for me. But you can try to set it to 0 and see the result.

Good job - it’s not trivial to write such a functions…

It seems someone from Mikrotik staff was thinking in the same way. :smiley:
But I was thinking on generic snmp table and how to match the same row index number in different table columns. :laughing: