Hello all,

as mentioned in an another thread I have started a log shipping scenario
where there's an origin, a slave and a log shipping slave. The setup went
smoothly and everything worked when the logs produced by the normal
subscriber were applied to the log shipping slave. The problem is that
inevitably the replication to the log shipping slave stops. The reason is
that the syncs in the logfiles have gaps:

test2_log=3D# select * from _test2.sl_setsync_offline ;
 ssy_setid | ssy_seqno |         ssy_synctime
-----------+-----------+-------------------------------
         1 |       523 | 2007-07-20 15:19:54.493855+03
(1 row)


$ psql -f slony1_log_1_00000000000000000523.sql test2_log
START TRANSACTION
psql:slony1_log_1_00000000000000000523.sql:6: ERROR:  Slony-I: set 1 is on
sync 523, this archive log expects 522
ROLLBACK
VACUUM

$ psql -f slony1_log_1_00000000000000000524.sql test2_log
START TRANSACTION
psql:slony1_log_1_00000000000000000524.sql:5: ERROR:  Slony-I: set 1 is on
sync 523, this archive log expects 521
ROLLBACK
VACUUM

$ psql -f slony1_log_1_00000000000000000525.sql test2_log
START TRANSACTION
psql:slony1_log_1_00000000000000000525.sql:6: ERROR:  Slony-I: set 1 is on
sync 523, this archive log expects 524
ROLLBACK
VACUUM

Something has gone wrong between 524 and 525. In fact, 524 file reveals
that:

$ cat slony1_log_1_00000000000000000524.sql
-- Slony-I log shipping archive
-- Node 1, Event 524
start transaction;

select "_test2".setsyncTracking_offline(1, '521', '524', '2007-07-20 15:19:
58.812135');
-- end of log archiving header
------------------------------------------------------------------
-- start of Slony-I data
------------------------------------------------------------------

------------------------------------------------------------------
-- End Of Archive Log
------------------------------------------------------------------
commit;
vacuum analyze "_test2".sl_setsync_offline;

To my eye it seems that file 524 is trying to advance the sync number from
521 to 524, while it has already been incremented to 523 in the previous
file. Why does the 524 file try to advance the sync number in such a way?

This is slony 1.2.10 and pgsql 8.2.4 if it matters.

Regards

MP
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.slony.info/pipermail/slony1-general/attachments/20070720/=
38b1d970/attachment.htm