on 08-26-2013 8:42 PM
Hi,
The Landscape is as below :
Application Server (CI) - Host1
Message Server & ERS - Virtual Host on Host2 OR Host3 ( Cluster Envt - node failover ).
DB - Virtual Host on Host2
Host1 and Host2 are physical servers. OS is Windows.
We are using Solman 7.1 SP8 and its a Cluster Environment.
Now my question is when I use the On the Fly Agent Scenario as per the Agent Setup Strategy Document we run Solman_Setup --> Managed System --> hosts --> Select Hosts and run the configuration. Enable the "On the Fly Scenario" and here it automatically searches and selects the Instances running on the particular physical Host.
However, this is limited to the Instances which are already running on this server.
For Ex: MS & ERS are currently running on Host2
I run the "On the Fly Agent" setup for Host2 and these instances are read perfectly.
But when the node failover happens how does the Agent on Host3 find these instances since when the "on the fly" setup was run for Host3 these instances were not operational there.
If the above is correct then how should we maintain the different instances running on multiple hosts ? Do we add them manually by mapping on all the physical hosts where the instance could run ( primary & secondary ).
If not then please correct me.
Thanks,
Regards,
Kaustubh.
ES & MRS are running on host2. To have Agent on the fly to be working, install agent on physical host or virtual host and not on logical host. Once you setup agent on the fly on host2 you can select/exclude the logical hostnames.
To have failover to happen you also need to install an agent on failover node and perform agent on the fly here also.
Now agent on fly setup is done in both the primary and failover node. once the system moves to failover node you transfer all the logical hosts from primary to failover. Agent is started and monitoring is active.
Thanks,
Rohan
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Rohan,
I am asking when the failover of a particular instance happens to Failover, and the Managed Systems --> Hosts configurations ( On the fly configurations ) were run on that host when this particular instance was not running on it, then in that case how does/should/how the Agent dynamically creates the "agent process" for the instance.
I hope my question is clear and understood.
Regards,
Kaustubh.
Hi,
Actually the whole purpose of on the fly agent ( so called dia process) is to reduce the number of DIA agent and its foot prints in memory.
if you install single DIA agent with on the fly feature, this automatically creates multiple on the fly agent process for each and every instance in the server associated with the logical host names. This is also work for certain scenario that if your DB failover in host2 , then corresponding DB On the fly agent process replicate from Host 3. Rest your ERS and MS everything smooth from Host2, when your entire physical host fails then al the agent on fly moves from host 2 to host 3.
Regarding different instances running on different host, if you have HA environment, still you can define the agent on the fly.
you can get all the needed screenshots as mentioned link given by Roman above.
Please check.
Thanks
Jansi
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Jansi,
My question is :
Regards,
Kaustubh.
Please read the following wiki Diagnostics Agent and HA Support. Hope it helps you to understand Agents On-the-fly in HA context.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi,
Is there any advise on the above ?
Regards,
Kaustubh.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
89 | |
10 | |
9 | |
9 | |
9 | |
6 | |
6 | |
5 | |
5 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.