Processing attributes syntax

Processing attributes are used to control how files are accessed by the client.

Table 1 describes processing attributes. Defaults are underlined in this format. You can override the default processing attributes on the mount command or file processing commands.

Table 1. Processing attributes
Processing Attribute	Description
attrtimeout(n) The time (in seconds) that the data set remains allocated after a lookup or getattr server operation. The default value of n is 120. The value of n can range from 1 to 32,767 (9 hours, 6 minutes, and 7 seconds). Note: The attrtimeout value is normally greater than the readtimeout or writetimeout values. With NFS version 2 and version 3 protocols, the lookup operation searches for a file in the current directory. If it finds the file, lookup returns information on the file's attributes and a file handle pointing to the file. With NFS version 4, neither the file's attributes nor the file handle are returned. The file handle is saved by the server and used as an anchor for accessing the file. When using the NFS version 4 protocol, the attrtimeout value should be set to a value less than or equal to the lease time. Otherwise, it is possible for performance problems to occur when attempting to access MVS data sets. With NFS version 2 and version 3 protocols, the data set may be closed/deallocated before the timeout value has been reached if the data set has been requested by another application if the delegation site attribute or modify operator command V4DELG=on is specified. noattrtimeout The data set is not deallocated after a lookup or getattr operation. For more information, see Timeout attributes.
binary Indicates that the data set is processed between the client and server using binary format and no data conversion occurs between ASCII and EBCDIC formats. text Converts the contents in the data set between EBCDIC and ASCII formats. Use this format to share text data between clients and z/OS applications. In text mode, the following attributes apply only to z/OS MVS data sets: blankstrip and noblankstrip. (See the entry for blankstrip in this table.) End-of-line terminators (cr, crlf, lf, lfcr, or noeol) are used to indicate the MVS logical record boundary. (See the entry for lf in this table. See Text processing mode for rules of coding EOL terminators by the z/OS NFS server. See the xlat attribute in this table for customized EBCDIC-ASCII tables.)
blankstrip With text mode, strips trailing blanks from the end of each record of a fixed-length text file when the file is read. Trailing blanks pad the end of each file or record when a text file is written. noblankstrip Does not strip trailing blanks from the end of fixed-length records when a fixed-length text file is read. Does not pad records when writing a text file. The file must be of the correct size or an I/O error is reported to the client. For information about the text attribute, see the entry for binary in this table. This attribute does not apply to z/OS UNIX files.
With text mode, use one of the following end-of-line specifiers. cr Carriage Return is the end-of-line terminator. crlf Carriage Return followed by Line Feed is the end-of-line terminator (standard DOS). lf Line Feed is the end-of-line terminator (standard AIX® or UNIX). lfcr Line Feed followed by Carriage Return is the end-of-line terminator. noeol No end-of-line terminator. For information about the text attribute, see the entry for binary in this table. This attribute does not apply to z/OS UNIX files.
cln_ccsid(n) Specifies the coded character set identifier (CCSID) for the remote mounted file system (NFS client) when text is being translated. The default value of n is 819 (ISO 8859-1 ASCII).
*convserv(technique)* Specifies the conversion technique-search-order that Unicode Services will use for specified srv_ccsid(x) and cln_ccsid(x) code pages. Technique consists of up to five technique-characters corresponding to the available techniques: R, E, C, L and M. See z/OS Unicode Services User's Guide and Reference for detailed descriptions on these conversion techniques. NFS version 4 protocol (NFSv4) differs from NFSv2 and NFSv3 protocol in handling single to multiple byte conversion. Therefore, the technique-search-order specified in the convserv() attribute should consider the effects of the NFS protocol being used. SeeCreating the conversion environment for Unicode Services for further details. The default value of technique is LRE.
executebiton Sets the execute permission bits in user, group, and other (as reported with the ls AIX or UNIX command) for a mount point's files. Use when storing executable or shell scripts on the z/OS system. This option can only be overridden on a mount point basis — not at a command level. The executebiton attribute does not apply to z/OS UNIX files and can only be used with the mount command. executebitoff Does not set execute bits in user, group, and other for the mount point's files. This value is normally used in the site file.
extlink Specifies the use of the external link command to create, process, and delete a symbolic link to an MVS data set. The extlink attribute is used with the following commands. ln -s Create a symbolic link to an MVS data set. ls -l Display the attributes and contents of the symbolic link. rm Delete the symbolic link. The extlink attribute only applies to z/OS UNIX file objects. If extlink is not specified then it is not displayed by the showattr client enabling command.
fastfilesize Specifies to get the file size from SPF statistics, if it exists, for direct data sets, PDSs, and non-system-managed data sets. nofastfilesize Specifies to read the entire file or member to get the file size for direct data sets, PDSs, PDSEs, and non-system-managed data sets. Using the nofastfilesize attribute might cause a noticeable delay when first accessing very large data sets. These attributes apply to MVS data sets, but do not apply to z/OS UNIX files. For more information, see Using fastfilesize to avoid read-for-size.
fileextmap Enables file extension mapping. The fileextmap attribute can be specified at the file command level for the client platforms that support passing of attributes. See File extension mapping for related information. nofileextmap Disables file extension mapping.
mapleaddot Enables mapping of a single leading "." from a client file name to a legal leading "$" on z/OS. The mapleaddot attribute should normally be enabled for access by AIX and UNIX clients. nomapleaddot Disables mapping of a single leading "." from a client to a leading "$" on z/OS. These attributes do not apply to z/OS UNIX files.
maplower Enables mapping of lower case file names to upper case when accessing files on z/OS, and back to lower case when sending to the network. This option should normally be enabled for access by AIX or UNIX clients. This option only affects file names (high-level qualifiers and user catalog aliases). nomaplower Disables mapping of lower case file names to upper case and back to lower case when using files on z/OS. Note: These attributes do not apply to z/OS UNIX files.
mapped The mapped attribute should be specified at the mount or site level when a mixed set of data types is to be processed under a single mount point. The determination of whether the data is to be processed as text or binary depends on the rules that are established in the specified side file. See File extension mapping for related information. If a file extension is not mapped to text or to binary using the side files, then the data will be processed according to what has been specified at the mount or site level (binary or text). If binary or text is specified at the file command level, that overrides the mapped specification. If mapped is not specified then it is not displayed by the showattr client enabling command.
mvsmnt It is highly recommended that the mvsmnt processing attribute be specified on all NFS Version 4 user mount commands issued to the z/OS NFS server for three reasons. In NFS Version 4, mount requests are passed to the server in the form of a PUTROOTFH operation followed by a sequence of lookup operations. The mvsmnt processing attribute indicates to the z/OS NFS server that the associated lookup operation is emulating a mount procedure and causes the z/OS NFS server to write the mount point to the mount handle database (MHDB), so the z/OS NFS server can automatically recover the mount point during a server restart. Without the specification of the mvsmnt attribute, the z/OS NFS server must rely on the mount being restored via the FHEXPIRED error mount recovery process between the server and client after the server restart. Note: Some clients may not be able to recover from an FHEXPIRED error. See Implicit prefix support restrictions for an important note about the recovery of mount points after z/OS NFS server restart when there has been a change to the prefix site attributes. If the mvsmnt, or any other, processing attribute is not specified, then saf checking may be disabled longer than desired due to the z/OS server's inability to detect the end of the mount and the beginning of other access requests. However, access is allowed only to information necessary for the completion of mount related processing, and any non-mount related operations are always processed with full SAF enforcement. This behavior is required to avoid requiring an mvslogin to be issued prior to mount processing. The EXPORT file may be used to restrict access to file systems, regardless of the specification of any processing attribute. Note: The mvsmnt attribute must not be specified when mounting a path that contains a symbolic link whose underlying real path can change between z/OS NFS server instances. The specification of the mvsmnt attribute will prevent the z/OS NFS server from recognizing the underlying real path change upon restart. For any LOOKUPs that do not specify MVSMNT, any processing attributes that may have been provided will be merged with any that were in effect for the LOOKUP parent directory before applying to this LOOKUP. For LOOKUPs that do specify MVSMNT, any other processing attributes provided will be merged with the site defaults before applying to this LOOKUP. MVSMNT cannot be specified for any LOOKUP where the parent directory was navigated to by a mount procedure or a result of an object that was already LOOKUP'ed with MVSMNT. This is to ensure that only a client system mount specifies MVSMNT. mvsmnt is not displayed by the showattr client enabling command.
overflow Specifies x37 detection support for PS/PDS MVS data sets for NFS Version 2, Version 3, and Version 4 WRITE operations. It allows ENOSPACE errors to be reported to the NFS Client in a timely manner and to avoid situations when the z/OS NFS Server closes a data set on timeout expiration basis with x37 abend which can not be propagated to the NFS Client. This option may also be activated on the MOUNT level. Default mode is no overflow detection, and is not displayed by the showattr client enabling command.. Note: PDSE and VSAM data sets are not supported.
nordrcache Specifies that the server should not stale the legacy (MVS z/OS conventional data) internal readdir cache if an addition is made to the directory. This causes the next READDIR operation to access the directory information from the Physical File System (PFS) rather than the internal readdir cache. The default value is nordrcache and is not displayed by the showattr client enabling command. The nordrcache attribute does not apply to z/OS UNIX files. When nordrcache is not specified, the addition of an entry to the legacy internal readdir cache will not be visible to the client until the next readdir cache timeout or a remove from that directory. When nordrcache is specified, the addition will be visible to the client by the subsequent READDIR, whether the readdir cache timeout has expired or not. This may impair performance because the directory list must be read from the Physical File System after any addition to the cached directory. When nordrcache is specified, if no changes are made to the internal readdir cache, the cache does remain available until the readdir cache timeout expires.
nordrverf Specifies not to perform cookie verifier checking for the NFS version 3 readdir and readdirplus procedures, and the NFS version 4 readdir procedure. Note: nordrverf does not provide consistency in the listing of a directory's content and may cause duplicate or omit entries when the directory is changing during the listing. rdrverf Specifies to perform cookie verifier checking for the NFS version 3 readdir and readdirplus procedures, and the NFS version 4 readdir procedure. Note: rdrverf provides consistency in the listing of a directory content and, as a result, may impact performance.
readtimeout(n) The readtimeout attribute specifies the amount of time in seconds before a data set is released after a read operation. The value of n can range from 1 to 32,767 (9 hours, 6 minutes, and 7 seconds). The default value of n is 90. The server closes the file when the file times out. The readtimeout attribute does not apply to z/OS UNIX files. Note: When using the NFS version 4 protocol, the readtimeout value should be set to a value less than or equal to the lease time. Otherwise, it is possible for performance problems to occur when attempting to access MVS data sets. noreadtimeout Specifies the data set is not deallocated after a read operation. For more information, see Timeout attributes.
The z/OS NFS server uses DFSMShsm to recall or delete migrated files. The action that the server takes against migrated files depends on which of the retrieve or noretrieve attributes is active. The retrieve and noretrieve attributes do not apply to z/OS UNIX files. retrieve When the retrieve attribute is active, the server will recall the migrated file if necessary, upon an NFS_LOOKUP request for the file, depending on the files status. The server may be able to obtain the migrated files attributes without recall (see Retrieve attributes for additional information). If not, the recall operation is started by the server. The server waits for the recall operation to complete if the file resides on DASD. If the file does not reside on DASD, the server does not wait for the recall operation to complete and returns a "device not available" message. You can attempt accessing the file again later when the recall is complete. retrieve(wait) When the retrieve(wait) attribute is active, the server waits for the recall to finish. retrieve(nowait) When the retrieve(nowait) attribute is active, the server does not wait for the recall to finish, and immediately returns a "device not available" message. You can attempt accessing the file again later when the recall is complete. noretrieve When the noretrieve attribute is active, the server does not recall the file, and returns "device not available" upon an NFS_LOOKUP, NFS_READ, or NFS_CREATE request for a migrated data set. For more information, see Retrieve attributes.
setownerroot Specifies that z/OS NFS server return root user authority as the owner of a z/OS MVS data set when the client is logged on as superuser. setownerroot does not grant root authority for a UID=0 user, instead see the mvslogin command or the root keyword in the exports file. The setownerroot attribute can only be used with the mount command and does not apply to z/OS UNIX files. setownernobody Specifies setting the user ID in a file's attributes to nobody (65534), for a superuser.
sidefile(dsname) Specifies the name of the data set that contains the rules for file extension mapping purposes. If a side file name is specified in the attributes data set, then it is used as the default side file for the NFS server. A user can also specify an additional side file name during a mount operation to be used along with the default. The mapping rules will first be searched in the side file specified during the mount command and then the default side file is searched. To allow file extension mapping, a side file name must be specified either as a default or in the mount command. The value of dsname is a fully-qualified MVS data set name without quotation marks. See File extension mapping for related information. If a side file name is not specified in the attribute file or in the mount command, sidefile() is displayed by the showattr client enabling command.
srv_ccsid(n) Specifies the coded character set identifier (CCSID) for the local mounted file system (NFS server) when a new file is being created. The srv_ccsid attribute has no effect on the translation of an existing file's data. The default value of n is 1047 (`LATIN OPEN SYSTEM EBCDIC`). If the srv_ccsid attribute is not specified, new z/OS UNIX files will continue to be created as untagged.
sync Specifies that data transmitted with the write request should be committed to nonvolatile media (for example, DASD) by the server immediately when received. async The user can alternatively specify the async processing attribute to get improved performance. When this attribute is specified, the data will be committed asynchronously. The sync\|async attribute only applies to z/OS UNIX file objects and only for the NFS version 2 protocol.
tag Specifies that the newly created files should receive a file tag. notag Specifies that the newly created files should not receive a file tag. The tag is set to 0x0000. See Native ASCII processing attributes and Table 1 for considerations when using the tag and notag processing attributes.
writetimeout(n,o) Specifies the amount of time n, in seconds, before a data set is released after a write operation and the amount of time o, in seconds, that the server will wait for data to arrive to complete a partial record before closing the data set. The value of n can range from 1 to 32,767 (9 hours, 6 minutes, and 7 seconds). The default value of n is 30. The value of o can range from n to 255 * n. The default value of o is 120. The server closes the file when the file times out. All cached buffers are forced to disk. Normally writetimeout values are kept short because write operations result in exclusive locking. However, for slow client machines with long pauses between writes, you should increase the writetimeout value. The server will use the o value to extend the writetimeout value for a data set processed in text or binary mode in the case of a partial record (no end-of-line terminator discovered in the record or RPC WRITE data was lost in the network) on a WRITE operation to delay file closing and wait for the record completion data to arrive, so that the server is able to correctly process the partial record. The writetimeout attribute does not apply to z/OS UNIX files. Note: When using the NFS version 4 protocol, the writetimeout value should be set to a value less than or equal to the lease time. Otherwise, it is possible for performance problems to occur when attempting to access MVS data sets. nowritetimeout Specifies that the data set is not deallocated after a write operation. For more information, see Timeout attributes.
xlat(member_name) Specifies how to override the installation default translation table during file processing. The member_name is the member name of the PDS or PDSE that contains the customized translation table. The system administrator defines this member name in the attribute data set, and PDS or PDSE in the startup procedure. The xlat attribute is ignored if specified on the command line. If a customized translation table is not specified in the attribute file or in the mount command, xlat() is displayed by the showattr client enabling command.
znfsclient Specifies that the NFS client is a z/OS NFS client. The z/OS NFS server uses this attribute to customize its response to the NFS client. Note: Do not specify this attribute. This attribute is automatically appended to the processing attribute string by the z/OS NFS client when it detects that it is sending the mount request (or the last LOOKUP for an NFS version 4 mount) to a z/OS NFS Server. znfsclient is not displayed by the showattr client enabling command.