On 8/28/19 11:36 AM, francis picabia wrote:
>
>
> The other possibility was it just didn't log the slower times, but I
> have seen larger numbers
> in the 10,000 ms range in the night when some backups and housekeeping
> happens.
>
> All log related settings:
> checkpoint_segments = 12
> logging_collector = on
> log_destination = 'stderr'
> log_directory = 'pg_log'
> log_filename = 'postgresql-%a.log'
> log_truncate_on_rotation = on
> log_rotation_age = 1440
> log_rotation_size = 0
> log_min_messages = info
> log_min_error_statement = debug1
> log_duration = on
> log_line_prefix = '<%t>'
You might also want to log:
log_connections = on
log_disconnections = on
And add to log_line_prefix some mix of:
%u User name
%p Process ID
%s Process start time stamp
>
> I know it does sound strange, but this is what we battled with for 2.5
> days until the light
> came on that the vacuum had been set to off on the target system during
> server migration.
>
>
--
Adrian Klaver
adrian.klaver@aklaver.com