High Performance Computing

Hi,

I am trying to install Solaris cluster with Solaris10 and Cluster suite 3.2 2/08. After going through customer configuration in "scinstall" utility, I am facing a problem.

At the time of selecting 2nd node's file system type, I am giving input as /globaldevices which is very much part of my 2nd node's partitions, I am getting error as,

What is the name of the file system? /globaldevices

Testing for "/globaldevices" on "solariscluster1" ... failed

scrcmd: RPC: Program not registered
Unable to check "/globaldevices" on "solariscluster1" due to failed remote command.

Please if anybody knows, let me know how to go ahead from here onwards?

Thanks,
Hitarth

/globaldevices (or the underlying device) has to be cluster unique. So each
cluster node has to use a different meta device.
Check that the hostnames are in both the /etc/hosts /etc/inet/ipnodes files in both nodes. Be sure you don't have some unprintable character in vfstab .
Try to get the install log from the system and post its messages

Hi,

Sorry for replying late as I was out of station for some days. Anyways, I am back and tried again with all suggestions but still I am facing problems. Let me tell you how I tried step by step.

(1) Installed Solaris10 on both node having "/globaldevices" as separate partition. (I haven't registered OS to SUN for some reason).
(2) Installed Solaris Cluster Suite 3.2 2/08 on both nodes.
(3) Make entries of both node in files like /etc/hosts,/etc/inet/ipnodes. (Please see attached file where I have given respective configurations what I did before moving ahead with "scinstall").
(4) ssh is enabled between two nodes and root user login is also available when accessing from root user.
(5) Both nodes have cross-cable connectivity for Private network and that physical ethernet card is not plumbed at startup as well!!!!
(6) After going ahead with "scinstall" custom installation from Node-1, I got stucked and not getting how to move ahead. (Attaching file for scinstall log from Node-1).

I think I had the same problem once setting up a sun cluster, the problem was caused by RPC which rejected connections not from localhost.
First try if you see the requested service from localhost (rpcinfo -p) then try to access the service from other host (seconde cluster node), use the rpcinfo command with options (-p host; -t ) if you can connect from localhost, but can't from remote your're home!

GOT IT Ha Ha:
these services have to be reconfigured to enable rpc queries from hosts other then localhost:

metad
metamedd
metamhd
scadmd
scrcmd
metacld

the command to reconfigure them:
# svccfg -s rpc/bind listprop config/local_only
config/local_only boolean true

# svccfg -s rpc/bind setprop config/local_only=false

# svccfg -s rpc/bind listprop config/local_only
config/local_only boolean false

# svcadm refresh svc:/network/rpc/bind


Try googling for rpc remote/ rpc allow remote/ rpc localhost... you'll find NOTHING.
Knowing right peaple is important these days ....

It was very nice to refresh cluster experience a bit. Thank you, the pleaseure was all mine

AFAIK, sun's docummentation tells nothing about this as well ...

Togr,

First, You are AWSOME and yes we have to catch RIGHT person these days to come out of any task if you are NEWBIE like me. I am catching you now!!!!

I have followed your suggested steps and tried again to build a two node cluster but nodes are not coming in cluster mode after rebooting by scinstall.

########## TWO NODE Cluster attempt ################
bash-3.00# clnode show
clnode: (C152734) This node is not in cluster mode.
#############################################

This messages is coming on both servers. I tried to create SINGLE node cluster on one node as well but after rebooting it, same above mentioned message!!!!

############## One Node Cluster Log ####################

bash-3.00# cat scinstall.log.5275

*** Establish Just the First Node of a New Cluster ***
Fri Oct 24 16:00:15 IST 2008


scinstall -ik -C hitarth -F -o -P task=quorum,state=INIT


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

Initializing cluster name to "hitarth" ... done
Initializing authentication options ... done


Setting the node ID for "solariscluster1" ... done (id=1)


Checking for global devices global file system ... done
Updating vfstab ... done

Updating nsswitch.conf ... done

Adding cluster node entries to /etc/inet/hosts ... done


Configuring IP multipathing groups ...done

Verifying that power management is NOT configured ... done
Unconfiguring power management ... done
/etc/power.conf has been renamed to /etc/power.conf.102408160153
Power management is incompatible with the HA goals of the cluster.
Please do not attempt to re-configure power management.

Ensure network routing is disabled ... done
Network routing has been disabled on this node by creating /etc/notrouter.
Having a cluster node act as a router is not supported by Sun Cluster.
Please do not re-enable network routing.

Rebooting ...
###################################################

P.S.- (1) Both nodes are connected with CROSSOVER cable for private network. (2) I haven't DISABLE auto quorum option while creating this cluster.

Please take out of this!!!!! I am getting frustrated now!!!!

Thanks