SUMMARY: SunCluster 3.2: LogicalHost resource not bringing back up aninterface ...

From: Noelle M. Vega <nmvega_at_computingarchitects.com>
Date: Thu May 24 2007 - 09:36:15 EDT
Hi:

For the problem I explained, its actually not (as told to me by Sun). IPMP handles
underlying problems with SC logicalHost VIPs within an IPMP group (moving
them as necessary from one interface to another - normal IPMP behavior).

But IPMP does not concern itself with the VIP's themselves as it does
with link level loss and with reachability to well known IP's on the
local subnet (e.g. default router; multicast address; etc.).

So the behavior I experienced below, is not incorrect.

Btw... Thank you Andreas for pointing out that SC3.2 support starts at
the H/W 11/06 version of Solaris 10.

Regards,
Noel Milton

==================


Hi:

Im running SunCluster 3.2 on Solaris 10 (06/06).

I have a one-node cluster up and running. The resource group I defined has,
as part of it, a LogicalHost defined as shown:

==================================================================
root# clrs show -g mq1-RG -p NetIfList,HostnameList mq1-lh-RES

=== Resources
===



Resource:                                       mq1-lh-RES

  --- Standard and extension
properties ---



  NetIfList:
ipmp_grp0@1

Class:                                         extension

Description:                                   List of IPMP groups on each
node
    Per-node:                                      False

Type:                                          stringarray
==================================================================


I can bring the resource group online and offline with each resource correspondingly
going on and offline as well, including the LogicalHost resource.

My questions is this...

With the resource group brought online, ce0:2 gets configured successfully (LogicalHost).
Fine. But when I, on purpose, do an "ifconfig ce0:2 down" or unplumb it, to hopefully
see the cluster bring it back up, an scstat shows the LogicalHost as online
even though ce0:2 remains down or unplumbed:

            Resource Name  Node Name                State          Status Message
            -------------  ---------                -----          --------------
  Resource: mq1-lh-RES         server22                 Online         Online - LogicalHostname online.

  Resource: mq1-hadisk-RES  server22                 Online         Online

  Resource: mq1-mq-RES       server22                 Online         Online - Service is online.


Anyone know why it doesn't bring it back up? I followed the docs to the letter.

TIA,
Noel Milton
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
Received on Thu May 24 09:36:30 2007

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:44:05 EST