[Bucardo-general] Strange log errors after upgrade Wheezy to Jessie

[Paul Theodoropoulos]

Recently upgrade our 'master' master servers from Debian Wheezy to 
Debian Jessie - which includes an upgrade of postgresql from 9.1 to 9.4.

After rebuilding Bucardo for the new os, bucardo runs seemingly fine, 
except for some peculiar errors in the logs, which I'm not sure how 
serious they are to the actual replication. I have a 'pre' server and a 
production server which were upgraded within a few days of each other, 
and they exhibit different errors, but both within VAC:

server 1, preproduction:

(15077) [Mon Oct 17 12:48:35 2016] VAC New VAC daemon. PID=15077
(15077) [Mon Oct 17 12:48:35 2016] VAC Connected to database 
"local_trumgr_group" with backend PID of 15079
(12910) [Mon Oct 17 12:48:35 2016] MCP Created VAC 15077
(15077) [Mon Oct 17 12:48:36 2016] VAC Connected to database 
"remote_trumgr_group" with backend PID of 6490
(15077) [Mon Oct 17 12:48:36 2016] VAC Warning! VAC was killed at line 
7727: DBD::Pg::st pg_result failed: ERROR:  relation 
"bucardo.delta_bucardo_delta_public_unit_history_78" does not exist
LINE 1: DELETE FROM bucardo.delta_bucardo_delta_public_unit_history_...
                     ^
QUERY:  DELETE FROM bucardo.delta_bucardo_delta_public_unit_history_78 
USING (SELECT txntime AS tt FROM 
bucardo.track_bucardo_delta_public_unit_history_78 GROUP BY 1 HAVING 
COUNT(*) = 1) AS foo WHERE txntime = tt AND txntime < now() - interval 
'45 seconds'
CONTEXT:  PL/pgSQL function bucardo.bucardo_purge_delta_oid(text,oid) 
line 49 at EXECUTE statement
SQL statement "SELECT 
bucardo.bucardo_purge_delta_oid($1, myrec.tablename)"
PL/pgSQL function bucardo_purge_delta(text) line 12 at SQL statement at 
/usr/local/share/perl/5.20.2/Bucardo.pm line 7727.

and the production server:

(3081) [Mon Oct 17 12:36:48 2016] VAC New VAC daemon. PID=3081
(5536) [Mon Oct 17 12:36:48 2016] KID (trumgr_group_sync) Totals: 
deletes=1 inserts=2 conflicts=0
(3749) [Mon Oct 17 12:36:48 2016] MCP Created VAC 3081
(3081) [Mon Oct 17 12:36:49 2016] VAC Connected to database 
"remote_trumgr_group" with backend PID of 22267
(3081) [Mon Oct 17 12:36:50 2016] VAC Connected to database 
"remote_trumgr_main" with backend PID of 22268
(3081) [Mon Oct 17 12:36:50 2016] VAC Connected to database 
"local_trumgr_group" with backend PID of 3157
(3081) [Mon Oct 17 12:36:50 2016] VAC Connected to database 
"local_trumgr_main" with backend PID of 3167
(3081) [Mon Oct 17 12:36:50 2016] VAC Warning! VAC was killed at line 
7727: DBD::Pg::st pg_result failed: ERROR:  relation 
"bucardo.delta_bucardo_stage_public_notification_rules" does not exist
LINE 1: DELETE FROM bucardo.delta_bucardo_stage_public_notification_...
                     ^
QUERY:  DELETE FROM 
bucardo.delta_bucardo_stage_public_notification_rules USING (SELECT 
txntime AS tt FROM bucardo.track_bucardo_stage_public_notification_rules 
GROUP BY 1 HAVING COUNT(*) = 1) AS foo WHERE txntime = tt AND txntime < 
now() - interval '45 seconds'
CONTEXT:  PL/pgSQL function bucardo.bucardo_purge_delta_oid(text,oid) 
line 49 at EXECUTE statement
SQL statement "SELECT 
bucardo.bucardo_purge_delta_oid($1, myrec.tablename)"
PL/pgSQL function bucardo_purge_delta(text) line 12 at SQL statement at 
/usr/local/share/perl/5.20.2/Bucardo.pm line 7727.


Thoughts? These errors did not occur before the upgrades.

-- 
Paul Theodoropoulos
www.anastrophe.com