Configure ACE to Monitor CDS-IS

Sandeep Singh — Tue, 29 Aug 2017 12:03:09 GMT

Introduction

This document explains how to configure probes on ACE in order to monitor Service Router, Streamer, and Origin Server. By default, ACE will consider a host to be opertaional if it is able to resolve its IP address, meaning that the host will be perceived as up and running as long as it is replying to ARP request, even if the service is down. This article explains how to configure probes on ACE in order to monitor the following components:

Service Router

Streamer

Origin Server

The goal is to detect a failure of one physical device and take it out of the load-balancing pool, and generate an snmp trap. Assuming that we have proper connectivity, it is also possible to use the ACE to monitor those appliances even without using the ACE to loadbalance them.

SNMP Trigger

Refer to CISCO-ENHANCED-SLB-MIB and following link for details about how to trigger a trap when a host fails: MIB

http://www.cisco.com/en/US/docs/interfaces_modules/services_modules/ace/vA2_3_0/configuration/administration/guide/snmp.html#wp1204030

Configuration for Service Router

TCP Level Probing

One can probe the Service Router on tcp port 80 (HTTP), tcp port 1935 (RTMP), tcp port 554 (rtsp).

Step 1. Configure the probe

probe tcp TCP_80

interval 5 // probe fired ever 4 seconds

receive 2 // no complete HS after 2 seconds = 1 probe failed

faildetect 2 // two consecutive probe failure = host failed

passdetect interval 3 // once failed, need 3 success to consider host back up again

connection term forced // RST after success, no graceful shutdown

Step 2. Assign the probe(s) to serverfarm

serverfarm host CDE-CACHES

probe TCP_80

probe TCP_1935

rserver LB-SR1 80

inservice

rserver LB-SR2 80

inservice

Remark: by default the host fails if *one* probe fails. To fail the host only if *all* probes fail, use "fail-on-all"

Step 3. Verify the probe status and hosts status

switch/C2# sh probe TCP_80

probe : TCP_80

type : TCP

state : ACTIVE

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

port : 80 address : 0.0.0.0 addr type : -

interval : 4 pass intvl : 3 pass count : 3

fail count: 2 recv timeout: 10

--------------------- probe results --------------------

probe association probed-address probes failed passed health

------------------- ---------------+----------+----------+----------+-------

serverfarm : CDE-CACHES

real : LB-SR1[80]

10.1.22.4 21 0 21 SUCCESS

real : LB-SR2[80]

10.1.22.3 21 0 21 SUCCESS

switch/C2#

switch/C2# sh serverfarm CDE-CACHES

serverfarm : CDE-CACHES, type: HOST

total rservers : 2

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

----------connections-----------

real weight state current total failures

---+---------------------+------+------------+----------+----------+---------

rserver: LB-SR1

10.1.22.4:80 8 OPERATIONAL 0 3 9

rserver: LB-SR2

10.1.22.3:80 8 OPERATIONAL 0 0 6

switch/C2#

DNS

You can probe and check that the Service Router is able to resolve DNS (e.g. one service as shown here, or the hostname of one CDS-IS host)

probe dns DNS_PROBE

receive 2

domain ds1.cds.cisco.com

HTTP

For HTTP, you can probe the Service Router and see if we receive the 302 for a particular service. The probe below will be successful if receive any http reponse code comprised between 200 and 302

probe http SR_HTTP

interval 5

faildetect 2

receive 2

request method get url ds1.cds.cisco.com

expect status 200 302

connection term forced

Configuration for Streamer

You can configure exactly the same tcp and http probes for the streamers as for Service Router. The valid http response code for a streamer should be only 200 or 206 or 304.

Configuration for Origin Server

The Origin server has same probes as those for Streamer.

Related Information

Release Notes for Cisco Internet Streamer CDS 2.6.3

article Configure ACE to Monitor CDS-IS in Data Center and Cloud Knowledge Base

Configure ACE to Monitor CDS-IS

Introduction