HP Serviceguard Version A.11.20 Release Notes, April 2011

To provide the required redundancy for both networking and mass storage connectivity on servers
with fewer than three I/O slots, you may need to use multifunction I/O cards that contain both
networking and mass storage ports.
Token Ring and FDDI Obsolete
Serviceguard A.11.20 does not support Token Ring and FDDI technologies for the cluster heartbeat
and data networks (Serviceguard A.11.17 on HP-UX 11i v2 and A.11.16 on HP-UX 11i v1 were
the last versions that did). HP-UX 11i v3 does not support these two technologies.
The unsupported configurations include physical Token Ring and FDDI interfaces, Virtual LAN
(VLAN) interfaces over FDDI or Token Ring, and failover groups of Token Ring and FDDI interfaces
in the LAN Monitor Mode of the APA product.
Parallel SCSI Dual Cluster Lock Obsolete
You must use Fibre Channel connections for a dual cluster lock; you can no longer implement it in
a parallel SCSI configuration (as of Serviceguard A.11.18). See “Dual Lock Disk” in Chapter 3
of Managing Serviceguard for more information about dual cluster locks.
Parallel SCSI Not Supported for Lock LUN
The lock LUN functionality does not support parallel SCSI; you must use Fibre Channel for a lock
LUN. If you need to use parallel SCSI, use an LVM cluster lock disk, or a Quorum Server.
For more information about the cluster lock, see “Cluster Lock” in Chapter 3 of the latest edition
of Managing Serviceguard.
Cluster Name Restrictions
The following characters must not be used in the cluster name if you are using the Quorum Server:
at-sign (@), equal-sign (=), or-sign (|), semicolon (;).
These characters are deprecated, meaning that you should not use them, even if you are not using
the Quorum Server, because they will be illegal in a future Serviceguard release. Future releases
will require the cluster name to:
• Begin and end with an alphanumeric character
• Otherwise use only alphanumeric characters, or dot (.), hyphen (-), or underscore (_)
Optimizing Performance when Activating LVM Volume Groups
If a package activates a large number of volume groups, you can improve the package’s start-up
and shutdown performance by carefully tuning the concurrent_vgchange_operations
parameter in the package configuration file (or control script, for legacy packages).
Tune performance by increasing this parameter a little at a time and monitoring the effect on
performance at each step; stop increasing it, or reset it to a lower level, as soon as performance
starts to level off or decline.
Factors you need to take into account include the number of CPUs, the amount of available memory,
the HP-UX kernel settings for nfile and nproc, and the number and characteristics of other
packages that will be running on the node.
NOTE: Remember to do this exercise not only on the node on which the package will normally
run, but also on the node with the least resources in the cluster, as a failover or other unexpected
circumstances could result in that node running the package.
For more information, see the section “Optimizing for Large Numbers of Storage Units” in Chapter
6 of the latest edition of Managing Serviceguard (in the High Availability collection on
http://www.hp.com/go/hpux-serviceguard-docs) and the comments in the package configuration
file.
Announcements 11