Hi,

I have 2 computers and 2 databases in my replication set up.  Simple 1 mast=
er 1 slave set up, 2 sequences and 23 tables replicated id's from 1 - 25.  =
I am using Windows 2003 Server for both machines and Postgres 8.3.3.  I wro=
te a program between the two machines so that if the master restarts the pr=
evious slave will become the new master.  So it would drop his slony table =
and wait for the other computer to come back, drop his as well, then run th=
e slonik script to become the new master.  =


I have been testing this set up at my office and everything seemed to be ok=
.  I then put the programs on my live system which has the 2 computers seve=
ral kilometers apart.  I started to get this error:

ERROR -- Slony-I: setAddSequence_int(): sequence ID 1 has already been assi=
gned
STATEMENT -- select "_slony_db".setAddSequence_int(1, 1, "TABLENAME"."seq_i=
d1",")
NOTICE -- there is no transaction in progress

I think this means there is an entry in my sl_table or sl_seq table with an=
 id of 1?  I'm not sure how this happens.  Could it happen if one side is t=
rying to drop its slony table while the other is running the slonik script?=
  Anyways, let's say that is causing the error.  Fine.  What I don't unders=
tand is this.  To fix this, I tried dropping the slony table in replication=
 on both sides to drop all my slony tables.  Then reran my slonik script.  =
I got the same error.  I restarted slony-i in my services.  I got the same =
error.  I dropped my slony tables, restarted postgres as a service.  Reran =
my slonik script and *STILL* I got the same error.  The only thing that see=
med to work was dropping my entire database and restoring it from the ident=
ical one on the other server.  I also tried to delete the entry in my sl_se=
q table but it then said my seq was missing an id.  And then I couldn't dro=
p my slony-i replication entry in Replication.  I was then forced to drop m=
y database again.  =


What is also weird is that I then moved the boxes I had at the office to my=
 live system thinking it might be latency on the network because the machin=
es were so far apart.  I could not reproduce the problem.  Then I thought i=
t might be the database I was using was corruted so I backed it up and rest=
ored it on the boxes from the office, no problems again.  I'm puzzled.

So I guess my questions are:

1. What could be causing this problem?

2. How can I solve this without dropping my entire database which seems to =
be dangerous as one day I might corrupt both by accident and be well . . . =
screwed.  =


LW . . . Lawrence

_________________________________________________________________
More storage. Better anti-spam and antivirus protection. Hotmail makes it s=
imple.
http://go.microsoft.com/?linkid=3D9671357
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.slony.info/pipermail/slony1-general/attachments/20090808/=
3cd9c443/attachment.htm