Table of Contents
Advertisement
Chapter 2
HPSS Planning
Table 2-6 HPSS Dynamic Variables (Continued)(Subsystem Specific)
Variable
Avg. Text Overflows
Per Name Space
Object
Avg Length of Text
Overflows
Filesets per Name
Server
Avg. Bitfile
Segments Per Tape
Bitfile
Avg. Storage
Segments Per Disk
VV
Avg. Storage
Segments Per Disk
Alloc Rec
Avg. Storage
Segments Per Tape
Bitfile
Max BFS Storage
Segment
Checkpoints
Max Queued Bitfiles
Changing COS
Max Queued BFS
Storage Segment
Unlinks
Max Disk Bitfiles
Queued for
Migration
Max Disk Bitfiles
Queued for Purging
124
Description
A name-space object record can store a filename that is 23 characters long (the base
name, not the full pathname). If a filename is longer than 23 characters, a text overflow
record must be generated. Also, if a comment is attached to a name-space object, a text
overflow record must be created. This value defines, on average, how many text
overflow records will be created per name-space object. For example, if 5 in 100 name-
space objects will either have a filename longer than 23 characters or a comment
attached, a value of .05 would be entered.
This is the average length of each name space object which will exceed 23 characters.
This value in conjunction with "Avg Text Overflows Per Name Space Object" is used
to determine the amount of metadata in 'nstext' required to store all text overflows.
This value can significantly impact the size of the metadata estimate.
The number of filesets a Name Server will be managing.
The average number of bitfile segments created for each tape bitfile. When files are
migrated to tape from disk, there normally will be only one tape bitfile segment for the
file. However, if the file is partially modified, additional bitfile segments will be
migrated to tape. This value significantly impacts the overall metadata sizing estimate.
The average number of storage segments that will be created per disk virtual volume.
Since there are a maximum of 16,384 blocks per disk VV, and each storage segment
uses one or more VV blocks, the number must be less than or equal to 16,384. This
value can be determined by dividing the average disk VV size by the average storage
segment size.
Each disk allocation record stores information for up to 8 disk storage segments. If, on
average, 1 extra disk allocation record is required per 5 disk bitfiles, a value of 1.2
would be entered.
The average number of storage segments per tape bitfile. If files are only written to tape
through migration, the value can be 1, assuming the storage class characteristics have
been set up properly. If users are allowed to write directly to tape, the average should be
something higher than 1.
The maximum number of storage segment checkpoint records created by BFS. A value
of 500 should be sufficient for planning purposes.
The maximum number of bitfiles that have a pending class of service change. A value
of 1,000 should be sufficient for planning purposes.
The maximum number of storage segments that could be queued to be unlinked. A
value of 1,000 should be sufficient for planning purposes.
The maximum number of unmigrated bitfiles that may exist on disk at one time. The
value should generally equal Max Bitfiles on Disk.
The maximum number of migrated bitfiles that may exist on disk at one time. The
value should generally equal Max Bitfiles on Disk.
September 2002
HPSS Installation Guide
Release 4.5, Revision 2
Advertisement
Table of Contents
