Question on OracleSunCluster on Solaris guests

[vanaidu2004]

Hi All,

I'm installing Oracle Sun Cluster on Solaris 10 guests with below configuration.
Guest OS version: Solaris 10 9/10 s10x_u9wos_14a X86
Oracle Sun Cluster version: 3.2 2/08 (3.2U1)
Virtual Box version: 4.1.0

No issues found for installing guest OS and sun cluster softwares. When i try to configure the cluster it is failing with the below pasted error. I've assigned 4 network adapters for each guest OS and two of them are used to configure IPMP and remaining two are planned for cluster interconnect. I selected "typical" while configuring cluster.
As per the error i understand that some issue with network configuration but not sure what to change. I really appreciate if someone can help me to resolve the issue.
===================================================================================================================
Configuring "clhost3" ...

scrcmd -N clhost3 install -logfile /var/cluster/logs/install/scinstall.log.7280 -k -C lakki -F -T node=clhost4,node=clhost3,au
thtype=sys -w netaddr=172.16.0.0,netmask=255.255.248.0,maxnodes=64,maxprivatenets=10 -A trtype=dlpi,name=e1000g2 -A trtype=dlp
i,name=e1000g3 -B type=switch,name=switch1 -B type=switch,name=switch1 -m endpoint=:e1000g2,endpoint=switch1 -m endpoint=:e100
0g3,endpoint=switch1
scrconf: Failed to add connection point - already exists.
scinstall: Failed to initialize cluster configuration ("-B type=switch,name=switch1")

scinstall: scinstall did NOT complete successfully!

Checking device to use for global devices file system ... done

Initializing cluster name to "lakki" ... done
Initializing authentication options ... done
Initializing configuration for adapter "e1000g2" ... done
Initializing configuration for adapter "e1000g3" ... done
Initializing configuration for switch "switch1" ... done
Initializing configuration for switch "switch1" ... failed
==========================================================================================

Thanks in advance.

[dadd]

A little further along in the setup it will ask about auto discovery of the interfaces on the other node, default is "yes." Tell it "no" and go through the added steps, the answers seemed to be correct by default. For some reason it will put both interfaces on one node on the same switch. That should take care of that problem. The next one, if you are like me, is after the setup and configuration runs the other node will reboot and then crash. If you reboot the node you are running scinstall on that will crash also. Looking for an answer to that problem.

Cheers

[vanaidu2004]

I selected "Custom" installation and specified the network adapter names to use from other node(clhost3) for cluster interconnect. It configured the other node(clhost3) and rebooted it, as you said it's started crashing. But the node which i used to configure the cluster(clhost4) seems to be hung and it did not come out of the configuration setup. Usually it has to reboot to complete the setup.

I tried to change the network adapter type in virtual box setup. But all the 4 adapters are configured to use (Host-only adapter, 'VirtualBox Host-Only Ethernet Adapter'). Any further IDEAS!!!
====================================================================================================
>>> Cluster Transport Adapters and Cables <<<

You must configure the cluster transport adapters for each node in the
cluster. These are the adapters which attach to the private cluster
interconnect.

Select the first cluster transport adapter for "clhost4":

1) e1000g2
2) e1000g3
3) Other

Option: 1

Adapter "e1000g2" is an Ethernet adapter.

Searching for any unexpected network traffic on "e1000g2" ... done
Unexpected network traffic was seen on "e1000g2".
"e1000g2" may be cabled to a public network.

Do you want to use "e1000g2" anyway (yes/no) [no]? yes

The "dlpi" transport type will be set for this cluster.

Name of adapter on "clhost3" to which "e1000g2" is connected? e1000g2

Select the second cluster transport adapter for "clhost4":

1) e1000g2
2) e1000g3
3) Other

Option: 2

Adapter "e1000g3" is an Ethernet adapter.

Searching for any unexpected network traffic on "e1000g3" ... done
Verification completed. No traffic was detected over a 10 second
sample period.

Name of adapter on "clhost3" to which "e1000g3" is connected? e1000g3

Unused switch "switch1" will be ignored.
Unused switch "switch2" will be ignored.
Unused switch "switch1" will be ignored.
Unused switch "switch2" will be ignored.
====================================================================================================

[dadd]

The first node is waiting for the other node (clhost3) to come back up. After it sees that it will then configure itself and reboot. If you select "Failsafe" and fix the problems with clhost3 and reboot it may finish the install. You may then have the same problem with the first node. Still working on the solution to that. I decided to do some cloning as I was sick of reinstalling the whole thing everytime it crashed and I wanted to work with a clean system.

I had a simular problem with a hardware setup using 2 V120's and a StorEdge S1. I needed to change the scsi-iniator-id on one of the systems or they would start panicing and crash. Leaning to a SCSI problem within Virtual Box.



I set my first network adapter to "Bridged Adapter." The others I set to "Internal Network", I only used 3 cards. When I used other settings the system gives the error you have "Unexpected network traffic was seen on "e1000g2"." This allows me to use my host network to connect to other systems but the interconnect aren't bothered by the host network.

Here is my options:

1) Create a new cluster or add a cluster node

1) Create a new cluster

2) Custom

Name and node according to your system.

DES [NO]

Default address [YES]

Default netmask [YES]

Two PRIVATE NETWORKS [YES]

Use swithces [YES]

First switch [switch1]

Second switch [switch2]

1) e1000g1

Dedicated [yes]
Note here with internal networks you get no traffic.

Connected to switch [switch1]

Default port name "e1000g1" [yes]

2) e1000g2

Dedicated [yes]
You should have no traffic here also.

Connected to switch [switch2]

Default port name "e1000g2" [yes]

Is it okay to use autodiscovery for the other nodes no
***This where you don't allow the program to use auto. The answers seemed to be all correct after you give the node2 adapter names. You can do it without the switch as you did but should it ever work and you want to try a 3rd node you won't be able to, at least that is my understanding.***

for node "your.node.2.name"
Name of the first transport adapter e1000g1

Dedicated [yes]

Switch [switch1]

Default port name [yes]

Second transport adapter e1000g2

Dedicated [yes]

Switch [switch2]

Default port name [yes]

Finished with default answers. I used the default /globaldevices.

This time I took out all shared drives, I letf the SAS adapter, not SCSI, and the nodes both rebooted as expected, no errors. Next I shut down both nodes. Node2 had issues so I just did the power off from VB. I added a shared SAS drive to both nodes and started them up, node1 first. The system seemed to hang looking for node2 so I started that one up. Both systems came up without errors and "format" shows the shared disk.

I guess this is a success. Now to see how stable the thing is.

Good Luck.

Cheers

[vanaidu2004]

Thanks much for your detailed steps. I follwed the same steps and nodes are rebooted as expected and it worked fine. Let's see how it works further. Once again thanks for sharing your knowledge.