On z/OS V2.1? CMOD Works Just Fine With New Format zFS Aggregates

[Ed_Arnold]

History:   when zFS first came out, a recurring question was "Does CMOD work with zFS's?"

The answer was always yes.  Completely transparent.

With z/OS V2.1 a new, and better format of zFS aggregate has been made available.

I'm not going to write about it, all one would want to know is here:

http://www-304.ibm.com/support/knowledgecenter/SSLTBW_2.1.0/com.ibm.zos.v2r1.ioea700/ioea7p02999476.htm


Right now you can run the following JCL which simply does a brief display of information about each zFS aggregate on your system:


//AGGRINFO EXEC   PGM=IOEZADM,REGION=0M,
// PARM=('aggrinfo -long')             
//SYSPRINT DD     SYSOUT=H             
//STDOUT   DD     SYSOUT=H             
//STDERR   DD     SYSOUT=H             
//SYSUDUMP DD     SYSOUT=H             
//CEEDUMP  DD     SYSOUT=H             
//*                                     

This is output for my V9.0 product zFS before converting it to the new format:


ARS.ARSV900.PRODUCT.ZFS (R/W COMP): 119662 K free out of total 1080000
version 1.4                                                           
auditfid E2D4E2E5 D3F20598 0000                                       
         14957 free 8k blocks;          6 free 1K fragments           
          2160 K log file;             56 K filesystem table         
           160 K bitmap file                                         

Note the "version 1.4" in the output above, that's the old format.

I went into my BPXPRMxx and changed that mount to this:

MOUNT FILESYSTEM('ARS.ARSV900.PRODUCT.ZFS')
      MOUNTPOINT('/usr/lpp/ars/V9R0M0')   
      TYPE(ZFS)                           
      MODE(RDWR) PARM('CONVERTTOV5')       

All I did was add the PARM statement and IPLed.

Voila - now when I run the aggrinfo job I get:

ARS.ARSV900.PRODUCT.ZFS (R/W COMP): 119662 K free out of total 1080000
version 1.5                                                           
auditfid E2D4E2E5 D3F20598 0000                                       
converttov5                                                           
         14957 free 8k blocks;          6 free 1K fragments           
          2160 K log file;             56 K filesystem table         
           160 K bitmap file                                         

I have converted all of my product zFS's and all of my temp and cache zFS's and run a test load through my system. 

It works.

Ed

[rjchavez]

Great to know Ed!  Thanks for the info.  Note that you can also use this utility/command to convert the zFS without an IPL:

ioefsutl converttov5 -aggregate name -aggrversion_only [-verbose][-level][-help]

[Ed_Arnold]

To have all new aggregates automatically in V5 mode add this to IOEFSPRM:

format_aggrversion=4  Sets the default format for new zfs datasets to v4                 
format_aggrversion=5  Sets the default format for new zFS datasets to v5

There is no mass conversion utility, but there is something useful and that is the change_aggrversion_on_mount IOEFSPRM option.

With this option enabled zFS v4 datasets would be converted to v5 then next time they are mounted.

This can also be done with the converttov5 configuration option in IOEPRMxx or with CONVERTTOV5 on the mount command itself.

To maximize the conversion one could enable both change_aggrversion_on_mount and CONVERTTOV5.

Ref:

https://www.ibm.com/support/knowledgecenter/SSLTBW_2.2.0/com.ibm.zos.v2r2.ioea700/ioea7mt.htm

Ed

[Ed_Arnold]

A nice testimonial:

https://groups.google.com/forum/#!topic/bit.listserv.ibm-main/CLBgzhjJPQk

Ed