Stanza files

The input to a number of GPFS™ commands can be provided in a file organized in a stanza format.

A stanza is a series of whitespace-separated tokens that can span multiple lines. The beginning of a stanza is indicated by the presence of a stanza identifier as the first token on a line. Stanza identifiers consist of the % (percent sign) character, followed by a keyword, and ending with the : (colon) character. For example, %nsd: indicates the beginning of an NSD stanza.

A stanza identifier is followed by one or more stanza clauses describing different properties of the object. A stanza clause is defined as an Attribute=value pair.

Lines that start with the # (pound sign) character are considered comment lines and are ignored. Similarly, you can imbed inline comments following a stanza clause; all text after the # character is considered a comment.

The end of a stanza is indicated by one of the following:

a line that represents the beginning of a new stanza
a blank line
a non-comment line that does not contain the = character

GPFS recognizes a number of stanzas:

%nsd:: NSD stanza
%pdisk:: Physical disk stanza
%vdisk:: Virtual disk stanza
%da:: Declustered array stanza
%rg:: Recovery group stanza

The details are documented under the corresponding commands.

For more information about the GPFS Native RAID commands that use stanzas, see IBM Spectrum Scale RAID: Administration.

A stanza file can contain multiple types of stanzas. Commands that accept input in the form of stanza files expect the stanzas to be syntactically correct but will ignore stanzas that are not applicable to the particular command. Similarly, if a particular stanza clause has no meaning for a given command, it is ignored.

For backward compatibility, a stanza file may also contain traditional NSD descriptors, although their use is discouraged.

Here is what a stanza file may look like:

# Sample file containing two NSD stanzas

# Example for an NSD stanza with imbedded comments
%nsd: 	nsd=DATA5     # my name for this NSD
	device=/dev/hdisk5  # device name on node k145n05
 	usage=dataOnly
	# List of server nodes for this disk
	servers=k145n05,k145n06
	failureGroup=2
	pool=dataPoolA

# Example for a directly attached disk; most values are allowed to default
%nsd: 	nsd=DATA6   device=/dev/hdisk6   failureGroup=3