Metrocluster with Continuous Access XP P9000 for Linux Version B.01.00.00 Release Notes

Serviceguard Cluster and Package Configuration Via Serviceguard Manager or
System Management Homepage
Starting with HP Serviceguard Manager B.03.31.10 available in Serviceguard A.11.20.10, the
following Serviceguard Manager operations are supported in a Metrocluster environment:
• Configuration, monitoring and administration of Metrocluster modular packages.
However, the following restriction applies:
• A Metrocluster package configuration requires a system administrator to configure data
replication, which is not supported by Serviceguard Manager. This step is required to be
performed by the administrator outside of the HP Serviceguard Manager.
For more information, see HP Serviceguard A.11.20.10 for Linux Release Notes available at
http://www.hp.com/go/linux-serviceguard-docs and HP Serviceguard Manager B.03.31.10
online help system that is available with the product.
Timing Considerations
The Metrocluster with Continuous Access XP P9000 for Linux gets the device status from the P9000
or XP series array and/or synchronize the device group during package startup. The time taken
by package startup depends upon the number of disk devices configured. Packages with many
disk devices will take longer to start up than those with fewer devices.
Some delay in the package startup time occurs in those situations where resynchronization of data
from one side to another is required. For example, if the Metrocluster package fails over to SVOL
side and if there is a Continuous Access link failure, it performs SVOL takeover (SSWS). After
Continuous Access link recovery, if the Metrocluster package fails back to PVOL side, it
resynchronizes the data from the SVOL side to the PVOL side.
For Continuous Access Journal mode package, journal volumes in PVOL may contain a significant
amount of journal data to be transferred to SVOL, the package startup time may increase significantly
when the package fails over and waits for all of the journal data to be flushed.
The value of run_script_timeout in the package ASCII file should be set to no_timeout or
to a large value. This value should be enough to take into consideration the extra startup time that
is attributed to getting the status from the P9000 series disk array. When using asynchronous data
replication, the value has to be 360 seconds or more.
Additional Notes on Disks, Logical Volumes, and Filesystems
The following are the guidelines for the management of disks, LVM storage groups, logical volumes,
and filesystems in Metrocluster with Continuous Access XP P9000 for Linux packages:
• Metrocluster does not verifies the status of the host ports. Therefore, it is assumed that at least
one storage link is functional, otherwise the storage group activation fails.
• A given file system must not reside on more than one P9000 frame for either the PVOL or the
SVOL.
• A given storage group (LVM volume group) must not span across more than one P9000 frame
for either the PVOL or the SVOL.
• Each host must be connected to only the P9000 series disk arrays that contain either the PVOL
or the SVOL. A given host must not be connected to both the PVOL and the SVOL of a
Continuous Access pair except when using Remote Command Device configured over extended
SAN.
• When using the paircreate command to create PVOL/SVOL Continuous Access pairs,
specify the -c 15 switch to ensure the fastest resynchronization in order to reduce vulnerability
to a rolling disaster.
Timing Considerations 7