No you cannot avoid this.
The exit builds the added dictionaries in the path to allow for two things.
1) Which node, if using DPF, is sending the file to be archived.
2) To which Log Chain does this belong. Log chains are generated to a new
number every time you do a roll forward to a point in time. If you finished
the PIT in the middle of log S0000015.log of chain C000000 and also had
S0000016.log, S0000017.log in the chain and in the archive, then after the
roll forward and activate the db, you would generate S0000016.log and
S0000017.log in C0000001 which would send it to the that new directory
automatically when they archive.
That you don't step on your own showe laces as DB2 manages your archiving
properly.
Net is, you do not want to lose those added directories.
HTH, Pierre.
--
Pierre Saint-Jacques
SES Consultants Inc.
514-737-4515
<sa*******@gmail.coma écrit dans le message de news:
11**********************@p10g2000cwp.googlegroups. com...
We are currently using Userexit to archive DB2 logs. I updated the db
cfg as follows:
db2 update db cfg for capfpk02 using LOGARCHMETH1 DISK:/cfpkloc4/
archivedlogfiles/capfpk02
I assumed that the archived log files are going to the above
directory. But actually it is going to the following directory.
/cfpkloc4/archivedlogfiles/capfpk02/cfpkinst/CAPFPK02/NODE0000/
C0000000
It has added a lot of subdirectories. Can this be avoided? The high
availability guide says that the archived logs will go the directory
you specify.
Thanks for any help
naru Manglalath
