7.2.2 bug?
Seems it wants to run a redo entry that doesn't exist.
Not a big deal as it's a test environment only. It was recently
upgraded from 7.2.1 to 7.2.2. I'm wondering whether the person who did
the upgrade shutdown the daemon before installing.
FATAL 1: The database system is starting up
FATAL 1: The database system is starting up
DEBUG: database system is ready
DEBUG: server process (pid 9084) was terminated by signal 10
DEBUG: terminating any other active server processes
DEBUG: all server processes terminated; reinitializing shared memory
and semaphores
DEBUG: database system was interrupted at 2002-09-03 13:54:33 EDT
DEBUG: checkpoint record is at 0/1E1F1D90
DEBUG: redo record is at 0/1E1F1D90; undo record is at 0/0; shutdown
TRUE
DEBUG: next transaction id: 320415; next oid: 488052
DEBUG: database system was not properly shut down; automatic recovery
in progress
DEBUG: ReadRecord: record with zero length at 0/1E1F1DD0
DEBUG: redo is not required
FATAL 1: The database system is starting up
FATAL 1: The database system is starting up
DEBUG: database system is ready
DEBUG: server process (pid 9097) was terminated by signal 10
DEBUG: terminating any other active server processes
DEBUG: all server processes terminated; reinitializing shared memory
and semaphores
DEBUG: database system was interrupted at 2002-09-03 13:54:37 EDT
DEBUG: checkpoint record is at 0/1E1F1DD0
DEBUG: redo record is at 0/1E1F1DD0; undo record is at 0/0; shutdown
TRUE
DEBUG: next transaction id: 320415; next oid: 488052
DEBUG: database system was not properly shut down; automatic recovery
in progress
DEBUG: ReadRecord: record with zero length at 0/1E1F1E10
DEBUG: redo is not required
FATAL 1: The database system is starting up
Rod Taylor <rbt@zort.ca> writes:
DEBUG: server process (pid 9097) was terminated by signal 10
Could we have a backtrace from that core dump?
AFAICT it's getting through the WAL redo just fine, so the problem
is (probably) not what you think.
regards, tom lane
On Tue, 2002-09-03 at 16:42, Tom Lane wrote:
Rod Taylor <rbt@zort.ca> writes:
DEBUG: server process (pid 9097) was terminated by signal 10
Could we have a backtrace from that core dump?
AFAICT it's getting through the WAL redo just fine, so the problem
is (probably) not what you think.
Took me a while, but I eventually figured out that they changed the
NAMEDATALEN in the old version, and didn't match it in the new one.
So the error is exactly what is expected -- memory allocation errors.
Rod Taylor <rbt@zort.ca> writes:
Took me a while, but I eventually figured out that they changed the
NAMEDATALEN in the old version, and didn't match it in the new one.
Grumble. It occured to us to store NAMEDATALEN in pg_control in 7.3,
but 7.2 doesn't have that defense. Sorry bout that...
regards, tom lane