On 8/31/2004 12:36 AM, Ben wrote:
> Well, there are "idle in transaction" processes, but they're on other  
> databases than the two in question.
> 
> That shouldn't mess up slony, should it?

If you can show me an easy way how to tell if a given XID in progress 
will affect a particular databases committed status when it commits 
eventually in the future, then no. Until you can, the answer is yes, it 
does and should.

All PostgreSQL databases under one postmaster share the same XID space. 
There is no way I know to determine what database the connection belongs 
to that is executing any given XID.

Get rid of these idiot-java-grad-student-trial-project-autocommit-
feature-attempt hacks in your professional server environment, and slony 
will work fine.


Jan

> 
> On Aug 30, 2004, at 10:29 AM, Jan Wieck wrote:
> 
>> On 8/30/2004 1:10 PM, Ben wrote:
>>
>>> Highly doubtful. The process is pretty much exactly what is on the  
>>> intro documentation, and goes like so:
>>> export CLUSTERNAME=slony_test
>>> export MASTERDBNAME=bench
>>> export SLAVEDBNAME=bench2
>>> export MASTERHOST=localhost
>>> export SLAVEHOST=localhost
>>> export REPLICATIONUSER=postgres
>>> export DBUSER=bench
>>> createdb -O $DBUSER -h $MASTERHOST $MASTERDBNAME
>>> createdb -O $DBUSER -h $SLAVEHOST $SLAVEDBNAME
>>> echo "create table foo (id serial primary key);" | psql $MASTERDBNAME
>>> echo "create table foo (id serial primary key);" | psql $SLAVEDBNAME
>>> createlang plpgsql -h $MASTERHOST $MASTERDBNAME
>>> createlang plpgsql -h $MASTERHOST $SLAVEDBNAME
>>> ....and from there, I follow the intro docs at  
>>> http://gborg.postgresql.org/project/slony1/genpage.php?howto_basic.
>>> Well, with the change of fixing the typoo in the initial slonik  
>>> script. :)
>>> If there were some uncommitted transaction, how could I tell?
>>
>> ps ... look for any postgres processes that say "idle in transaction"  
>> in the ps -axl output.
>>
>>
>> Jan
>>
>>> On Mon, 30 Aug 2004, Jan Wieck wrote:
>>>> That isn't the error. The remote worker thread will not be there  
>>>> until there is a subscribed set to sync, not to copy. The warning is  
>>>> coming from a generic function that is called to wake one up if it  
>>>> exists.
>>>> I am pretty sure that you have some open connection. Are you for  
>>>> example connected to the DB with some other admin tool, like  
>>>> phppgadmin or the like? Is there any chance that some OTHER database  
>>>> under the same postmaster has connections that use bogus autocommit  
>>>> functionality?
>>> _______________________________________________
>>> Slony1-general mailing list
>>> Slony1-general at gborg.postgresql.org
>>> http://gborg.postgresql.org/mailman/listinfo/slony1-general
>>
>>
>> --  
>> #====================================================================== 
>> #
>> # It's easier to get forgiveness for being wrong than for being right.  
>> #
>> # Let's break this rule - forgive me.                                   
>> #
>> #================================================== JanWieck at Yahoo.com  
>> #
> 
> _______________________________________________
> Slony1-general mailing list
> Slony1-general at gborg.postgresql.org
> http://gborg.postgresql.org/mailman/listinfo/slony1-general


-- 
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me.                                  #
#================================================== JanWieck at Yahoo.com #