AIX

Hello,

I was wondering if I have 3 nodes (A, B, C) all configured to startup with HACMP, but I would like to configure HACMP in such a way:

1) Node B should startup first. After the cluster successfully starts up and mounts all the filesystems, then

2) Node A, and Node C should startup !
How can I configure this ?

Any help would be appreciated.

AIX version 6
HACMP/ES 5.4

ThanksAIX

Don't know whether I got you correctly. I have never seen an admin who wanted his clusters to start up at os boot. In case you are new to HACMP you might not be in an ideal situation for making such a cluster design decision. You can configure HACMP to start up with the os but if you can take some piece of advice: don't do it. IMHO the potential problems by far outweigh the potential adavantages.

HACMP is very flexible here, it lets you script everything and you can control when a command is run by pre or post events or by the start/stop scripts. You could create an ssh key with empty password distribute it accordingly to the nodes you want to access and then script the start of the other clusters. Following your description a good place for the startup command might be at the bottom of the RG start script. If you mean to power on the servers themselves you need to script HMC or CSM commands to name two possibilities.

HACMP won't start with the OS boot.

What I meant to say is that after all the systems are up and running..I want to start HACMP but it should be in sequence.

I have 3 servers:
SAP Application services are running on 2 servers and Database is running on one of the servers..

So, the question was:

I want HACMP to start the Database server node B first, then it should join in the cluster the application server nodes A,C ;

It can be done by the start-stop script as you mentioned...But i was wondering if there any other way to do it in HACMP ?

Something called Startup Policy / home node online / parent - child dependencies / application monitoring ?


thanks

I'm glad to hear this.

So you need information about HACMP monitors and how to combine them with a run-time policy. A monitor watches out for the existance of a certain process and uses your scripts to react if the process is not there/disappears. Then there are the following run-time Resource Group dependencies available from HACMP 5.3 on:However from the information you supply it would be hard to give detailed information how to setup the RGs. I suggest you check with the cluster docs about how to configure process monitors and custom monitors. Secondly read about the difference between startup monitoring mode and long running mode or the combined monitor that will check both start and runtime. Only monitors and scripts you define/write will ensure that the DB/app will start up in the order you wish.
High Availability Cluster Multi-Processing (HACMP) publications - IBM e(logo)Server Cluster Information Center

thanks for the answer. I have checked the guide... If you know any simple guide or procedure anywhere on the internet please let me know...


thanks once again.