Re: Burned by SAVCHGOBJ

[Vern Hamberg <vhamberg@xxxxxxxxxxxxxxxxxxxxxxxxx>]

Yeah, some reaqlly odd things. We have a procedure that does a SAVCHGxxx 
every night, with the previous run's date and time for reference date. We 
save to SAVFs, then copy to IFS and compress with gzip before writing to 
CD. We do libraries individually. One lacuna I just noticed is, a new 
library is not saved with SAVLIB - objects in it get caught by the 
SAVCHGOBJ, and restoring might be a problem if the library does not exist. 
Not too much of a problem - recovery is a manual process for now.

One problem is that the reference date function is specified differently in 
all the various SAVCHGxxx commands - Yuk!

But we've been able to recover from 2 inadvertent DLTLIB's

Dies BRMS allow saves to SAVF?

Regards

Vern

At 02:43 PM 3/25/2003 -0500, you wrote:
> The problem is most likely caused when you use the Reference Date Parm as 
> *SAVLIB, this causes a few problems.  One of the reasons we always have 
> our clients use the actual date of the full save of the library.
> If you create a new library and have never saved it using a SAVLIB type 
> command, then SAVCHGOBJ will not backup that library on each nightly 
> save.  Also, if someone Saves the library with update history *YES, then 
> that becomes the new reference point of backup whats been saved since that 
> date/time.
>
> SAVCHGOBJ is a great tool when you understand all the quirks that come 
> along with it. Implemented correctly, and the right procedures for 
> restoration, SAVCHGOBJ can work.
>
> What IBM is missing is the command SaveEverythingSinceMyLastFullSave 
> command.  Why would that be so difficult?
>
>    Pete