ACE: probe timers

[gavin han]

Hi,

I've general question about ACE probe timers. I've following probe setup:

probe https probe:1061

  port 1061

  interval 34

  passdetect interval 17

  open 1

ACE# sh probe probe:1061detail

probe       : probe:1061

type        : HTTPS

state       : ACTIVE

description :

----------------------------------------------

   port      : 1061   address     : 0.0.0.0         addr type  : -

   interval  : 34      pass intvl  : 17              pass count : 3

   fail count: 3       recv timeout: 10

===

for above probe: when ACE will declare the server as down? will it declare it down after (17*3+34) 85 seconds or it will declare it down after 115 seconds (added recv timeout=secs 3 times = 30 seconds).

please help.

========

we did a test and bought down the server manually. ACE declared the server down after 91 seconds (from the time when server was brought down).

[Krishnendu AR]

Even I got a few doubt regarding ACE probes

probe tcp TCP-10080

  port 10080

  interval 10

  faildetect 2

  passdetect interval 60

  passdetect count 2

What does each line in the above config stands for?

Krishna

[Daniel Arrondo Ostiz]

Hi Gavin, Krishna,

The explanation for all these parameters can be found in the health monitoring section of the configuration guide (

http://www.cisco.com/en/US/docs/interfaces_modules/services_modules/ace/vA2_3_0/configuration/slb/guide/probe.html#wp1031040)

Below are the definitions quoted from the guide:

Interval:

The time interval between probes is the frequency  that the ACE sends probes to a server marked as passed. You can change  the time interval between probes by using the interval command

Faildetect:

Before the ACE marks a server as failed, it must  detect that probes have failed a consecutive number of times. By  default, when three consecutive probes have failed, the ACE marks the  server as failed. You can configure this number of failed probes by  using the faildetect command

Passdetect interval/count:

To configure the time interval after which the ACE  sends a probe to a failed server and the number of consecutive  successful probes required to mark the server as passed, use the passdetect command.

So, to summarize, taking Gavin's configuration as example. A server failure would be detected in a time between 78 seconds (2x34 +10) and 112 (3x34 +10). Once it's down, it will become operational between 34 (2x17) and 51 (3x17) seconds after it comes back up.

I hope this helps

Daniel