Обсуждение: Re: Can we exclude errors of some particular text in Postgres logfile?
Hi Team, Can we exclude errors of a particular type from the postgres log file ? For example: ERROR: duplicate key value violates unique constraint "table_uniq" DETAIL: Key (column1, column2, column3)=(value0, value1, 1) already exists. If yes, how can we exclude this errors from the pg_log files. Suggest me an approach. Thanks in Advance. Regards, Pavan -- Sent from: http://www.postgresql-archive.org/PostgreSQL-admin-f2076596.html
On Sat, Sep 1, 2018 at 6:14 AM, pavan95 <pavan.postgresdba@gmail.com> wrote:
Hi Team,
Can we exclude errors of a particular type from the postgres log file ?
For example:
ERROR: duplicate key value violates unique constraint "table_uniq"
DETAIL: Key (column1, column2, column3)=(value0, value1, 1) already exists.
If yes, how can we exclude this errors from the pg_log files. Suggest me an
approach.
Thanks in Advance.
Regards,
Pavan
--
Sent from: http://www.postgresql-archive.org/PostgreSQL-admin-f2076596. html
I suggest upgrading to at least PostgreSQL 9.5 if you haven't already and look into the INSERT...ON CONFLICT feature (also called UPSERT). With it you can define what happens if an insert would violate a defined constraint and instead update the given row, do nothing or a wide variety of other operations.
If this doesn't fit the model that is causing your duplicate key errors, you may want to look into using a function to do your writes instead and catch the given error in an exception block and then just ignore it.
Keith
Re: Can we exclude errors of some particular text in Postgres log file?
От
"David G. Johnston"
Дата:
On Saturday, September 1, 2018, pavan95 <pavan.postgresdba@gmail.com> wrote:
Hi Team,
Can we exclude errors of a particular type from the postgres log file ?
No, it is not possible to apply textual filtering during log generation.
David J.
Hi Keith, I'm getting this error due to logical replication configured. And this error is occurring on the Subscriber Side. In my point of view, Logical Replication worker process is continuously checking and trying to insert each row right from the start even if that particular row has been already present in the table. And that's the reason of this continuous occurring of the error. One small doubt is if a table A has 10 records(which were inserted on the publisher) and all the 10 records were replicated to the subscriber. Now when I insert the 11'th record will it replicate only the last record or try to replicate(or check if all rows have been replicated) all the rows of the table A right from the first?? Please confirm.. If it does the latter(try to replicate all rows) then the above error will arise as table A has replica identity(Primary key). >>If this doesn't fit the model that is causing your duplicate key errors, you may want to look into using a function to do your writes instead and catch the given error in an exception block and then just ignore it. Could you please provide me an example of accomplishing it!! Thanks in Advance. Regards, Pavan -- Sent from: http://www.postgresql-archive.org/PostgreSQL-admin-f2076596.html