HP Serviceguard Version A.11.18 Release Notes, September 2008

NIC teaming (bonding) as in the fourth case listed in the warning text (VMware
ESX Server).
JAGag46475 (SR8606494542): cmquerycl -f line output may display incorrect VG disk
device files
• Problem: If a shared volume group used different device file names in /etc/lvmtab
on different nodes, the output displayed by cmquerycl -f line showed the
device files for all nodes as configured for the first node referenced in the -n
<node> option of cmquerycl. This could result in Serviceguard Manager creating
a cluster configuration file which had the wrong device file names for the cluster
lock volume group for some of the cluster nodes.
• Resolution: The code has been modified to display the correct device file names for
each node.
JAGag47195 (SR8606495273): Improper log message from cmmodnet on a system
on which IPv6 is not enabled
• Problem: When IPv6 was not enabled on a system, using cmmodnet to add an IPv6
address to an interface produced the following incorrect error message:
<address> might already be configured as a heartbeat IP or stationary IP address. Or it does
not belong to subnet <subnet>.
The real problem was that IPv6 was not enabled; cmmodnet did not check whether
IPv6 was enabled on the system before trying to add the address.
• Resolution: cmmodnet now checks whether IPv6 is enabled on the system before
trying to add the address.
JAGag48606 (SR8606496688): cmmigratepkg missing PEV and external script
permission
• Problem: The-e option of cmmigratepkg should generate Package Environment
Variables (PEVs) but it did not. (See “Migrating Legacy to Modular Packages”
(page 38) for more information about cmmigratepkg.)
In addition, the permissions on external scripts output by cmmigratepkg were
wrongly set to 555, giving everyone read and execute permission; 744 is correct,
giving only read permission to users other than the owner, who must be the root
user.
• Resolution: The code has been modified to correct these problems.
QXCR1000594093: cmcheckconf failed to detect cluster lock error
• Problem: cmcheckconf failed to report an error if a lock disk configured in the
cluster configuration file was missing or its volume group was not activated.
cmapplyconf does detect and report the error.
• Resolution: cmcheckconf and cmapplyconf both now behave correctly.
Fixed in This Version 81