Re: Timing of SAVLIB/RSTLIB vs CPYLIB

Consider a CLRLIB followed by a SAVRSTLIB as it may be faster again
than doing a Save and then a Restore.

Also keep in mind that CPYLIB may not get everything because as Vern
pointed out it uses CRTDUPOBJ under the covers - not every object can
be duplicated.

On Sat, Sep 11, 2010 at 12:19 AM, Vern Hamberg <vhamberg@xxxxxxxxxxx> wrote:

 CPYLIB uses CRTDUPOBJ for every object. It has to determine the
correct order for object dependencies, then it uses CRTDUPOBJ on each.
RSTLIB, esp. if you have saved access paths, will be faster, because the
SAV/RST code path is optimized for throughput.

Vern

On 9/9/2010 7:52 PM, Lennon_s_j@xxxxxxxxxxx wrote:

Curiosity question.

Each night in the wee hours we create a test environment from a
production library, thus:
      SAVLIB PROD01 to QTEMP/SAVF
      CLRLIB TST01
      RSTLIB QTEMPSAVF to TST01

It runs in about 41-45 minutes.

Last night I had to do a (occasional) copy of T01 to T02.  I did this:
      CLRLIB T02
      CPYLIB T01 to T02

It took 45 minutes.  I figured creating T02 this way would be quite a
bit faster than creating T01, but it wasn’t.

T01 was created about 3-3:45 am.  T02 was created 4:30-5:15 am.  There
is really nothing else running at these times.

Any thoughts on why the CPYLIB approach is not faster?   CPYLIB did seem
to do a lot of index rebuilding (observed in a test during the day.)
Maybe SAVLIB/RSTLIB is smarter about indexes and/or does more in parallel.

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.