Here's how to reproduce PostgreSQL dumping core.
I had a master/slave setup working OK with 8.1.3 running on both machines.
I upgraded to 1.2.6 on the master (./configure --with-perltools, make,
make install).
I then ran:
# slonik_update_nodes | slonik
<stdin>:5: loading of file /usr/local/pgsql/share//slony1_funcs.sql:
PGRES_FATAL_ERROR ERROR: Slonik version: 1.2.6 != Slony-I version in PG
build 1.2.2
ERROR: Slonik version: 1.2.6 != Slony-I version in PG build 1.2.2
########
I eventually realized I needed to also first update the slave, which I
did, using the same method:
Then I got a different result:
# slonik_update_nodes | slonik
<stdin>:5: commit transaction; - server closed the connection unexpectedly
This probably means the server terminated abnormally
before or while processing the request.
########
The logs on the slave showed it dumped a postgres.core file. Repeating
the last step didn't help. :)
Any tips on how I can get things back to stable state? I could provide
or analyze the core file if it would help.
Mark
_______________________________________________
Slony1-general mailing list
[email protected]
http://gborg.postgresql.org/mailman/listinfo/slony1-general
