On 10/22/2004 2:24 PM, Christopher Browne wrote:

> Matthew Wild <M.Wild at rl.ac.uk> writes:
>> On Fri, 22 Oct 2004, Ed L. wrote:
>>
>>> On Friday October 22 2004 8:20, Matthew Wild wrote:
>>> >
>>> > I'm trying to get slony working on a Tru64 master running postgres 7.3.7
>>> > replicating to a Linux slave running postgres 7.4.5 and have been held up
>>> > by the threadsafe-libpq problem. Is there a patch available for 7.3.7?
>>> > Would going to 7.4.5 help, or does that still need patching?
>>> 
>>> What specific indicators of a threadsafe-libpq problem are you seeing?  Can 
>>> you post error message or the like so others can tell if they're 
>>> experiencing the same thing?
>>>
>>
>> I get this continually repeating on the master node.
>>
>> ERROR remoteListenThread_2: "select ev_origin, ev_seqno, ev_timestamp,
>> ev_minxid, ev_maxxid, ev_xip, ev_t ype, ev_data1, ev_data2, ev_data3,
>> ev_data4, ev_data5, ev_data6, ev_data7, ev_data8 from "_w
>> dc_digisondes".sl_event e where (e.ev_origin = '2' and e.ev_seqno > '0')  
>> order by e.ev_origin, e.ev_seqno" - could not r eceive data from server:
>> Successful
>>
>> The slave is sitting there reporting nothing after this
>>
>> CONFIG storeSubscribe: sub_set=1 sub_provider=1 sub_forward='f'
> 
> I have seen the following message on Solaris when there was a mismatch
> vis-a-vis linking to the proper threaded libraries:
> 
> DEBUG1 remoteListenThread_1: connected to 'host=host004 dbname=pgbenchrep user=postgres port=5432'
> ERROR  remoteListenThread_1: "select ev_origin, ev_seqno, ev_timestamp,        ev_minxid, ev_maxxid, ev_xip,        ev_type,        ev_data1, ev_data2,        ev_data3, ev_data4,        ev_data5, ev_data6,        ev_data7, ev_data8 from "_pgbenchtest".sl_event e where (e.ev_origin = '1' and e.ev_seqno > '1') order by e.ev_origin, e.ev_seqno" - could not receive data from server: Operation now in progress
> 
> You sure seem to have hit the threading issue.
> 
> I'm not sure what the solution will be, but it's useful to at least
> track down where the error lies...

Usually the threading issue manifests itself in an "error=0" message 
during copy_set(). An easy way to test if it is a configuration problem 
or a general slony issue is to run one of the src/ducttape tests.


Jan

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