Shulou(Shulou.com)05/31 Report--

What this article shares with you is the analysis of the actual case of AIX HACMP cluster switching test. The editor thinks it is very practical, so I share it with you. I hope you can get something after reading this article. Let's take a look at it with the editor.

In order to verify the stability of AIX HACMP cluster system and ensure the continuity and reliability of its application, it is decided to carry out cluster switching test on Friday night. The following is a summary and record of the current process for future reference and filing.

System environment: AIX 5.3

Database: DB2 V8.2

Storage: IBM DS4700, configuring shared storage for two nodes

Schematic diagram of AIX HACMP principle:

HA Test 1: start and shut down

Ensure that VG is varyoff on both sides in the current system

# smitty clstart starts HA

Tail-f / tmp/hacmp.out look at the startup output

Tail-f / tmp/cm.log read the error message

Netstat-I and-in to see if the IP address becomes a service address

Lsvg-o check to see if VG is varyon on both machines

Check whether the corresponding SRC is up or not (2 or 3 SRC should be displayed).

Such as: Subsystem Group PID Status

ClsmuxpdES cluster 213488 active

ClstrmgrES cluster 238062 active

ClinfoES cluster 229656 active

# smitty clstop shuts down HA

Graceful (normal) / takeover (takeover) / force (mandatory, IP does not change back to boot)

Graceful is chosen here.

Tail-f / tmp/hacmp.out look at the shutdown output message

Tail-f / tmp/cm.log read the error message

Netstat-I and-in to see if the IP address becomes a boot address

Please check c-g cluster to see if the corresponding SRC is turned off (should not be displayed)

Using Cluster Snapshot, you can save the current Cluster configuration and restore it later by direct Apply.

HA Test 2: native IP takeover

Make sure that the HA of both machines has started normally.

Unplug the network cable from the network card associated with the host service IP.

Run on the host:

Netstat-I and-in

At this point, the standby network card will take over the service IP and the original service network card will correspond to the standby IP.

The standby machine does not have any reaction.

Plug back the unplugged network cable from the host and run:

Netstat-I and-in

At this time, the host does not have any recovery action, and the original standby network card still corresponds to service IP and the original service network.

The card still corresponds to standby IP, and even if HA is closed, it will not change back.

To restore the Nic to its original corresponding IP, run:

Smitty mktcpip

Select the network card (en0, en1,...) and enter without any changes.

HA Test 3: resource takeover

Make sure that the HA of both machines has started normally.

The simulation host terminates abnormally:

Method 1: run on the host:

Select takeover for smitty clstop shutdown mode

This kind of test is to take the initiative to tell the other party that it needs to take over, and it does not really simulate the fault.

Perform this step of testing before performing a real fault simulation.

Method 2: run halt-Q on the host to shut down immediately (the best method).

Method 3: unplug all network cables and heartbeats of the host at the same time.

Method 4: the mainframe is powered off (not recommended).

Run on the standby:

Netstat-I and-in

If the slave also has a service Nic (in cascading, that is, mutual takeover configuration), then the slave

The standby network card will take over the service IP of the host

If the slave does not have a service network card (in rotating, that is, standby configuration), then the boot network of the slave

The card will take over the service IP of the host.

When the slave has a svc address:

This is because the resource group will hang on the slave. If it does not hang up, it will run on the slave:

Varyonvg datavg

Mount-a (or hang one by one)

The FS defined in resource group should automatically mount on the standby.

Finally, test whether the application is normal or not, and complete the cluster switching test report.

The above is how to analyze the actual case of AIX HACMP cluster switching test. The editor believes that there are some knowledge points that we may see or use in our daily work. I hope you can learn more from this article. For more details, please follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.