As promised, here are the comparison statistics. It was an interesting 
excercise. I used a 1.12 system with 2 locals of 49500 tracks each, about 21% 
in use in the first dump, 14% in the second. 2GB central, 2 logical cps.

Dump options in the first dump (essentially 'dump everything'):
  CSA ALSO LSQA, SP(229:230,236:237,249) IN ASID(ALL)                         
  ALSO DATASPACES OF ASID(ALL)                                                
  ALSO PAGETABLES OF DATASPACES                                               
  ALSO SP(ALL) IN ASID(ALL)                                                   
  ALSO HIGH VIRTUAL IN ASID(ALL)                                              

Dump options in the second dump ('my options'):
  CSA ALSO LSQA, SP(229:230,236:237,249) IN ASID(ALL)                         
  ALSO RANGE(ALL) IN ASID(ALL)                                                
  ALSO DATASPACES OF 
ASID('CONSOLE','*MASTER*','RASP','OMVS','JES2','SMSPDSE','SMSPDSE1')
  ALSO SP(ALL) IN ASID(SYSKEY)                                                
  ALSO PAGETABLES OF DATASPACES                                               

Set the next to a fixed font, it appears in columns, first dump stats, then 
second dump stats:
Total Dump Statistics                                       
Elapsed time                      00:02:42.98       00:02:09.53
CPU busy percentage               4                 4
Average output data rate          14.52 MB/s        16.26 MB/s
Address space real pages          231,068           218,489
Data space real pages             206,406           217,841
High virtual real pages           47,350            40,163
Address space immediate aux pages 8                 2
Data space immediate aux pages    128               0
Address space deferred aux pages  165,950           112,167
Data space deferred aux pages     29,659            8,408
High virtual deferred aux pages   28,457            615

IARPFT Data Space Dump Statistics
Elapsed time                      00:00:02.96       00:00:03.60
CPU busy percentage               50                50
Average output data rate          12.49 MB/s        10.24 MB/s

In Real Virtual Dump Statistics                             
Elapsed time                      00:00:30.01       00:00:53.89
CPU busy percentage               10                5
Average output data rate          53.03 MB/s        28.99 MB/s
Address space real pages          231,068           218,489
Data space real pages             206,406           217,841
High virtual real pages           47,350            40,163

In Use Real Dump Statistics                                 
CPU busy percentage               9                 5
Average output data rate          25.89 MB/s        13.65 MB/s

Paged Out Virtual Dump Statistics                           
Elapsed time                      00:02:08.71       00:01:09.83
CPU busy percentage               1                 1
Average output data rate          5.15 MB/s         5.43 MB/s
Address space immediate aux pages 8                 2
Data space immediate aux pages    128               0
Address space deferred aux pages  165,950           112,167
Data space deferred aux pages     29,659            8,408
High virtual deferred aux pages   28,457            615
Single page read rate             6.64 MB/s         0.00 MB/s

Available Real Dump Statistics                               
Elapsed time                      00:00:00.71       00:00:01.20
CPU busy percentage               22                15
Average output data rate          85.23 MB/s        95.48 MB/s

What came as a surprise to me was that even with 'my options' quite a bit was 
read from the page datasets. Must be the range(all) in asid(all), plus we don't 
use much high virtual since we don't have enough page data sets.
Total elapsed time is only marginally longer when dumping everything, and the 
page data sets weren't as full and there wasn't as much dumped in the 'my 
options' case. I also have the sneaking suspicion that selecting things instead 
of dumping everything may take longer.

As far as I am concerned, 'my options' had been approved before and tested (in 
a 6GB real lpar it took about 10 minutes to dump a 1.8. system and 15 minutes 
before that for the operators to locate the right book containing the procedure 
and then do the procedure as described), so using autoipl/sadump, there will be 
less time spent on the sadump. I'll go in future with the 'dump everything' 
options.

Thanks to Jim for pushing me in the right direction. (Now, If only I knew what 
'Extra Buffers Batch Could Use     161,899' means... and if there is something 
I can do about that.....)

Best regards, Barbara

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Reply via email to