Hello,
I have this scenario:
1. 8.2.3 production server (32 bit) with WAL archive active
2. 8.2.17 backup server (32bit running on 64bit) that should keep
updating hourly the db cluster based on WAL files received from
production server
The problem is that after the first recovery the backup server has a
new timeline (2 instead of 1), so future recovery doesn't work as the
WAL files from the production server still have timeline id no 1.
What should I do?
1. modify postgresql in order to bypass creating a new timeline id
after each recovery, thus allowing me to start - recover - stop hourly
2. do something else (but what)?
I modified src/bin/pg_resetxlog/pg_resetxlog.c in order to save (-l
0x1,0xYY,0xZZ) the timeline but pg_control is invalid as there is no
checkpoint valid:
LOG: invalid primary checkpoint record
LOG: invalid secondary checkpoint link in control file
PANIC: could not locate a valid checkpoint record
Should I try to create a valid checkpoint in pg_resetxlog.c? (how to do that?)
Thank you
--
Best regards,
Dragos Moinescu