ADABCK and ASSO-Container

magsbg · September 26, 2005, 7:33pm

I defined a second ASSO-Container with the same size as the first one, a few days ago.
Since this the duration of ADABCK decreased from about 26 to 13 minutes.

Is there really such a strong connex between the performance of ADABCK
and the size of the ASSO-Containers or is it only a lucky coincidence ?

Matthias1 · September 27, 2005, 12:54pm

Are the ASSO-Containers located on two different disk drives?

magsbg · September 27, 2005, 1:11pm

No, both containers are on the same disk drive.

system · September 27, 2005, 3:23pm

Is this 2nd ASSO really used already, or just defined and “there” ?

Matthias1 · September 27, 2005, 4:45pm

Next questions (beside the one of WoWi):

Is the new BCK-Output-File smaller than the old one?
Is the Block-Size of the new ASSO-Container smaller than the old one?

IIRC, ADABCK stores only used Blocks into the BCK001-File. So it would be possible, that an ASSO with smaller Blocks produces less data to write.

magsbg · September 28, 2005, 2:24pm

The 2nd ASSO is really in use (I defined ASSO2 because ASSO1 was almost full).

The size of the BCK_Output is the same (it grew a little bit meanwhile).
The Block-Size of ASSO1 and ASSO2 is identical.

system · September 30, 2005, 11:39pm

Did you reorder or reload the database / files after adding ASSO2 ?

magsbg · October 3, 2005, 12:58am

Nothing at all.
ADABAS was running, when response 77 appeared - so I added ASSO2 without stopping the Database. No reorder or reload was started.