Mark,
>The total time of the outage from hang to CICS regions up an functional
>was 34 minutes which I know is a local record (not one that I am anxious
>to try and beat!). :-)
>Not bad considering this LPAR had 22G of real storage. The SADUMP
>performance with striping is amazing!
Can you provide a number how long dumping of real storage took? (Which options?)
I was able to test sadump on a 6GB real lpar, and the outage time from IPLing
sadump to IPLing MVS again was less than 10 minutes (dsntype=large to two
volumes):
8:18:50 IPL sadump
8:21:29 AMD095I VIRTUAL DUMP 15% COMPLETED. TOTAL MEGABYTES DUMPED: 8:25:02
End of sadump
8:28:08 Re-IPL 'Specify System parms' (and this only took so long because it
took me a while to recognize the end of sadump).
I keep getting postponed to test sadump on the big production lpar (well, 10G
is big for us) including their IMS/DB2 recovery. I want some numbers in case we
ever need an sadump in production to say: 'it'll only be 15 minutes longer, but
we'll have a dump to provide IBM'.
On a smaller lpar where I recently took several sadumps the actual dumping
process appeared shorter than when I took an svcdump (2.5GB real, the svcdump
always had to include OMVS and all dataspaces and the SMSPDSE/1 stuff). The
capture phase lasted 'forever'.
We also found out the hard way that when you need any OMVS stuff in the sadump
(like file information for open files), you'd better specify
ALSO DATASPACES OF ASID('OMVS')
Traditionally, we also dump all of RASPs dataspaces.
Regards, Barbara Nitz
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
