Serviceguard Version A.11.16 Release Notes, 2nd Edition, September 2004

Serviceguard Version A.11.16 Release Notes, Second Edition
Patches and Fixes in this Version
Chapter 1 43
From: Disks which do not have IDs cannot be included in
the topology description. Use pvcreate(1M) to initialize
a disk for LVM or, use vxdiskadm(1M) to initialize a disk
for VxVM.
To: Disks were discovered which are not in use by either
LVM or VxVM. Use pvcreate(1M) to initialize a disk for
LVM or, use vxdiskadm(1M) to initialize a disk for VxVM.
JAGae80291 Wrong order in vxdg deport for cluster node
shutdown
• What was the problem? Wrong order in vxdg deport for cluster node
shutdown in /sbin/init.d/cmcluster: It first deports the disk
groups and then halts the cluster nodes.
Since packages might still be running and file systems might still be
mounted when a shutdown command is issued, the deport commands
will fail. During shutdown, the cmcluster.init deports vxvm disk
group before halting the node. Deport can fail sometimes because the
package using it might be still running and file systems still
mounted. Sometimes, this can impact shutdown.
• What was the resolution? Fix is to move vxdg deport to after
cmhaltnode in halt_cluster_node.
JAGae69654servicefail fast occurs even on single node clusters. -
better error msg
• What was the problem? This is not a defect but a designed behavior.
Since VxVM-CVM-pkg uses FAILFAST option for both service and
node, once it fails, the node will TOC.
• What was the resolution? Logging messages are added to notify
customer when FAILFAST is enabled and that node can TOC when
package fails, even if it’s the only node in the cluster.
JAGae64683: System Multi-Node (SMN) packages aren’t handled
correctly by cmsnmpd
• What was the problem? The cmsnmpd does not send traps when a
SMN package is started, because the owner=0 in the Serviceguard
event is considered unowned. When a SMN goes down, the