Re: VOLSTATS on tapes

On 17-Feb-2014 10:23 -0800, Jeff Crosby wrote:

In looking at the info here

http://www.ibm.com/support/knowledgecenter/ssw_i5_71/cl/prterrlog.htm
(this is a beta site)

regarding the PRTERRLOG command, the help on the VOLSTAT parm says
(among other things):

"Note: ENDOPT(*UNLOAD) must be specified during the SAVE operation
to generate volume statistics at the completion of the tape
operation."

If this is this saying what I think it's saying, then:
If your SAVxxx command is ENDOPT(*REWIND) because you're going to
programmatically check that the object(s) made it to tape, or
ENDOPT(*LEAVE) because you've got more to add to the tape then the
volume stats are not updated? What if you're doing a Save 21 or
equivalent? Then you always have multiple SAVxxx commands following
each other.
Or does it mean simply that the stats are not updated until you
ultimately *UNLOAD the tape? What if you didn't unload by SW, but
pressed the eject button on the tape drive?

My impression is that the documentation, as often is the case, is trying to be succinct to avoid having given a long expansive dialogue to try to explain more broadly; at the expense of being somewhat misleading to those who think beyond the most direct perspective. Thus I would go with the latter. My inference regarding the statistics, is that the intention of the docs is to suggest that the statistics will not be updated until the Unload operation is performed for that volume, irrespective the origin being from CHKTAP ENDOPT(*UNLOAD), a SAVxxx ENDOPT(*UNLOAD), or any other method [though possibly only normal vs abnormal paths] effecting that unload... And the hardware unload I expect would be similarly registered\processed as a normal path, although a power outage or other device failure might not have the LIC Tape properly dump the statistics to the logs. The hardware eject delayed until the system was powered\powering off [if the pwrdwn does not effect an implied eject] I suppose might lose the statistics [although a design could ensure otherwise].

FWiW: The same note is presented for the PRTERRLOG CL command documentation in the IBM i 7.1 InfoCenter [and in the help text; both should be the same]. That is presumably the source for the information made available in that much slower link; i.e. apparently an unnecessary indirection to obtain and present within a frame, the Print Error Log (PRTERRLOG) CL command documentation found here:
<http://pic.dhe.ibm.com/infocenter/iseries/v7r1m0/topic/cl/prterrlog.htm>