Michael Crozier <crozierm at conducivetech.com> writes:

>> ERROR  remoteListenThread_2: "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 "_slony_regress1".sl_event e where (e.ev_origin = '2' and e.ev_seqno >
>> '0') order by e.ev_origin, e.ev_seqno" - could not receive data from
>> server: Error 0
>>
>> I do not see anything obviously wrong and the query succeeds from
>> psql. There are no errors in the postgresql logs.
>
> I could not determine why slon compiled with 7.3.10 had communcation
> problems, but I was able to get test 1 to complete by using the slon
> that was compiled against 8.0.5.
>
> This jogs my memory... I believe I did the same when I initially
> used slony to upgrade from 7.3 to 7.4.  While I would feel uncertain
> using this configuration permanently, it should suffice for upgrade
> purposes.  As I understand it, Postgresql 7.3 is no longer supported
> anyways.

7.3.12 just got released, so that's certainly an ambiguous thing :-).

At Afilias, we haven't had a 7.3 system around in quite a long time,
so it's not something *I'm* keen on putting terribly much effort into.  

If it's troublesome to use 1.1.5 with 7.3 on Solaris (and it may well
be a Solaris-ism that is the problem; I had no trouble running Slony-I
1.1.5 with PG 7.3.12 on Linux), then you could do worse than use 1.0.5
(which was pretty clean!) to upgrade from 7.3 to 7.4/8.0, and then
migrate to 1.1.5 to get to PG 8.1.

That's not perfectly ideal, but it's certainly better than saying
"tough luck!"

I don't think anyone's really keen on doing much work to figure out
7.3-related issues, so we're close to the point of "not supported."
-- 
(format nil "~S@~S" "cbbrowne" "ca.afilias.info")
<http://dba2.int.libertyrms.com/>
Christopher Browne
(416) 673-4124 (land)