topic Re: Changing Serviceguard Configuration in Operating System - HP-UX

Changing Serviceguard Configuration

???_185 — Mon, 20 Apr 2009 06:33:51 GMT

Hello.
I changed cluster configuration and met problem that I couldn't resolve.

* hpux 11.11 & serviceguard 11.16 on rp4440.

Existing cluster is configured with 1 stationary IP(subnet A), and 2 heartbeat IPs(subnet B, C: they connected to heartbeat dedicated switches with private IPs).
And 1 package controls a volume group and a RIP (subnet A).
Changes are add 1 more IP (subnet D), add it to package, and change default gateway from subnet A to subnet D.
(** subnet A and D are public addresses)

linkloop between new IPs(subnet D) are OK.
ie. NODE1 lan5,2 <--> NODE2 lan 5,2 all OK.
(on both nodes, lan2 is for redundancy.)

I modified netconf, hosts, cmclnodelist appropriately then did cmdeleteconf.

BUT cmcheckconf -C ascii failed with following message.
<<Please make sure that the proper security access is configured on node
NODE1 through either file-based access (pre-A.11.16 version) or role-based
access (version A.11.16 or higher) and/or that the host name lookup
on node NODE1 resolves the IP address correctly.
cmcheckconf: Failed to gather configuration information.>>>

identd(auth), hacl-* are running and LISTENING.
nslookup is OK.

I didn't solve the problem so removed new IPs and changed default gateway back to subnet A, then recompiled serviceguard without any errors.

What's wrong with me? How can I troubleshoot?
Any sugestions will be appreciated.

Thank you very much...

Re: Changing Serviceguard Configuration

Steven E. Protter — Mon, 20 Apr 2009 07:03:18 GMT

Shalom,

Changing hosts could have invalidated your cmnodelist. That could stop cluster communication.

Might need to do cmcheckconf/cmapplyconf, even cmquerycl before hand.

SEP

Re: Changing Serviceguard Configuration

Stephen Doud — Mon, 20 Apr 2009 10:24:08 GMT

Serviceguard employs 'identd' to qualify SG commands. identd must match the IP of the message to the hostname of one of the servers in the cluster. So hostname resolution must match any and all IPs that are assigned to NICs in /etc/rc.config.d/netconf to the server hostname.
The "Managing Serviceguard" manual describes how to configure /etc/hosts, but in short, this example demonstrates the principle:

15.145.162.131 gryf.uksr.hp.com gryf
10.8.0.131 gryf.uksr.hp.com gryf
10.8.1.131 gryf.uksr.hp.com gryf
15.145.162.132 sly.uksr.hp.com sly
10.8.0.132 sly.uksr.hp.com sly
10.8.1.132 sly.uksr.hp.com sly
15.145.162.150 bit.uksr.hp.com bit

Since you added a new subnet and IPs, insure they are aliased to the server hostname that supports them (on both nodes).

Re: Changing Serviceguard Configuration

Shardha — Tue, 28 Apr 2009 04:06:08 GMT

Dear

I belive you start from basics, as follows.

draw a diagram,
check and verify the connectivity.
recreate cmclnodelist file
run the querry script
modify the script, if requried.
run check script
finary run the apply script, if every thing is fine.

The diagram and nodelist file will be the key to success, as other specialist colleagues listed the point.

Shardha