For billing purposes we need to know how many bytes are written to tape. On a z/OS system using CA-1 their TMC has two fields Blocksize and Blockcount that give them that information. On a z/VM system using VM:Tape the TMC also has these fields but they do not get populated.
Our systems' main tape users are HiDRO, VM:Backup, VM:Archiver and VM:Spool. We also have SFS control file backups and EREP records but main users are CA products.
We have written a small pipe program that will mount a tape, read the tape, count the bytes and update the VM:Tape TMC but this ties up a tape drive and is not the way to go.
Do other z/VM system collect this data ? If so how ? Any suggestions ? Is there any support to request CA to modify their products to capture this data as is done for other OSes.
I know it is only part of the data you are trying to gather for billing, but VM:Archiver can produce accounting records for data storage on tape or dasd. They are put in the CP accounting stream. Information on the VM:Archiver accounting feature, including the record format is in the CA VM:Archiver Administrators wiki.
Since VM:Backup is their major user of tapes, this is what Graeme Moss has come up with:
1/ adding keyword report to all VMBACKUP job templates
1a/ add test to CHECKIT that VMBACKUP mdisk 1E1 is less than 80% full.
2/ add VMTAPE exit USERACNT which is called by the MOUNT command to write to user area of TMC a flag that this tape has to have byte field updated.
When RC report is to be generated
3/ if tape is flagged zero out byte count
4/ process all VMBACKUP keyword report files and update byte count and clear flag
5/ for tapes still flagged and CUSER is VMBACKUP write to alert report. (so we can add keyword report to template)
6/ for still flagged tapes including step 5 read tape and count bytes, update byte count and clear flag
7/ generate extract of TMC
Kitty has posted what we are investigating. A clarification of the above - CHECKIT is our daily health check.
A problem that has occurred with above process is that keyword reports are stored until a backup with same name is run then the report is replaced or the report is deleted by running the GETRPT command. There can be reports stored from years ago that have had their tapes scratched and reused many times but the report remains. This means added logic to filter the reports.
Having a batch system use the keyword reports has become messy. A better solution is for VM:Backup to update the TMC at the point it would write the report. This would also capture VM:Archiver usage.