Preparing for Cross-OS File Exchange Operations 3-11
Hitachi USP V Cross-OS File Exchange User’s Guide
– Defining RAW Devices. Those volumes which are to be used as
intermediate volumes and to be shared between open systems must be
defined as OPEN-3/8/9/K/E/L/M/V emulation type for RAID200/300,
and must be defined as “raw” devices from each host server. From the
open systems, there are no means to distinguish OPEN-3/8/9/K/E/L/M/V
for open system dedicated volumes from these FX volumes. Please make
sure not to confuse the usage on those volumes in the host systems.
• The operations below which create file systems on the intermediate
volumes must not be executed. Otherwise, information on the volume
may be destroyed and the volumes will become unusable as FX
volumes.
– Solaris: “newfs” command
– HP-UX: “pvcreate” command
– IBM AIX: creating a volume group
– Windows: formatting and creating a file system
– Digital UNIX/HP
®
Tru64 UNIX: “newfs” command
– SequeNT DYNIX/ptx
®
: creating a file system
– NCR
®
SVR4: creating a volume group
– Linux: “raw” command
Note for Microsoft
®
: Cluster Server: When installing FX devices in a
Microsoft Cluster Server (MSCS) environment, you must write
signatures on the FX volumes before configuring MSCS.
– The MSCS server cannot connect volumes which do not have
signatures.
– The volume on which a signature is written cannot be accessed from
another server.
– The volume on which a signature is written cannot be shared.
– Only the mainframe and the server which wrote the signature can
access the volume which has the signature.
– I/O time-out and I/O queue depth. Make sure to set the I/O time-out
and I/O queue depth values for the FX volumes as specified in the
applicable TagmaStore USP and NSC configuration guide.
– Partition size. Make sure to specify the correct partition size for the FX
volumes as specified in the applicable TagmaStore USP and NSC
configuration guide. If the partition size for -A or -B volumes is smaller
than the mainframe volume size, the open-system host may not be able
to access data to the end of the extent of these volumes.
