On Tue, Jun 21, 2011 at 1:40 PM, Steve Singer <ssinger at ca.afilias.info> wrote:
> On 11-06-21 01:35 PM, francescoboccacci at libero.it wrote:
>>
>
>> CONTEXT:  SQL statement "SELECT  "_sql_master_slave".subscribeSet_int( $1 ,
>> $2 ,  $3 ,
>> $4 ,  $5 )"
>>      PL/pgSQL function "subscribeset" line 77 at PERFORM
>>
>
> Slony normally prints this notice out as part of the subscription
> process. It is normal and does not indicate an error.

As a general rule, if the notification doesn't have the word "ERROR"
in it, then it isn't actually an error.

Slony reports various things that it does, so that an observing
administrator has some visibility of what it's up to.  In a number of
places, we RAISE NOTICE to indicate what's up.

Sometimes PostgreSQL automatically generates notices.  For instance,
if you create a table, and define a primary key, a notice will be
generated indicating extra actions taken.  For instance:

emp at localhost->  create table foo (id serial primary key);
NOTICE:  CREATE TABLE will create implicit sequence "foo_id_seq" for
serial column "foo.id"
NOTICE:  CREATE TABLE / PRIMARY KEY will create implicit index
"foo_pkey" for table "foo"
CREATE TABLE

It should be clear that those aren't errors.  When Slony generates
notices, they're not errors, either.

I think something is missing in the context information that was
reported; the information shown is neither an error message nor
notification of something that was done, but rather is merely
indicating where the thread of execution was when something else,
which presumably was output immediately before the CONTEXT line, was
reported.  I'd sure like to see what was reported BEFORE the
"CONTEXT:" line.

There has been some debugging code that wasn't very useful that has
gotten into the code base.  I don't always feel badly about that - it
can be useful to get that extra bit of output.  But there may be a
"RAISE NOTICE" statement that we could want to trim out.