Slow running R3szchk during Unicode System Copy

[Former Member]

Hi All,

we're running a homogeneous system copy as part of a unicode upgrade of our R/3 (4.7ext200) system. We've kicked off the SAPinst portion, and we're running the table size scans using R3szchk.

Unfortunatley this is really taking a long time....we killed it after 30 hours to further performance tune our database - but that hasn't made much of a difference. We'd really be expecting a complete run-time in the region of 10 - 12 hours at absolute worst.

Our hardware isn't being particularly well utilised - CPUs running at <40% and not much swapping in/out of physical memory going on.

Has anyone got any ideas as to why this might be happening - and what the problem might be?

All help appreciated.

Many thanks

Tim.

[Former Member]

Hi,

You may try to create the SPACE statistic with brconnect according to note 554031, otherwise R3szchk phase will need to select from DBA_SEGMENTS rather than read from DBSTATC which shold be a lot quicker.

Ole

[former_member269913]

I am facing the same problem,I locked the call to SAP they are looking into this issue. The only thing I remember is that I updated R3szchk bfore executing the sapinst, hoping that will be better. Seems like there is a glitch in the tool.

[Former Member]

you have to make sure all the database statistics have been updated - we did it via the SAPinst GUI. This speeded things up quite a bit. If you don't update the stats, R3szchk has to go and check the sizes out for itself, which is what takes the time.

Hope that's useful.

Cheers

Tim.