zabbix 1.8.2 / postgres stopped collecting data

[KHoos]

Today I noticed that zabbix 1.8.2 / postgresql 8.4.5 stopped gathering data. The zabbix_server log is filled with messages like:

 [select nextid from ids where nodeid=1 and table_name='events' and field_name='eventid']
  6942:20101103:120301.159 [Z3005] Query failed: [0] PGRES_FATAL_ERROR:ERROR:  current transaction is aborted, commands ignored until end of transaction block
 [select max(eventid) from events where eventid between 100000000000000 and 199999999999999]
  6942:20101103:120301.160 [Z3005] Query failed: [0] PGRES_FATAL_ERROR:ERROR:  current transaction is aborted, commands ignored until end of transaction block
 [insert into ids (nodeid,table_name,field_name,nextid) values (1,'events','eventid',100000000000000)]
  6942:20101103:120301.160 [Z3005] Query failed: [0] PGRES_FATAL_ERROR:ERROR:  current transaction is aborted, commands ignored until end of transaction block
 [update ids set nextid=nextid+1 where nodeid=1 and table_name='events' and field_name='eventid']
  6942:20101103:120301.160 [Z3005] Query failed: [0] PGRES_FATAL_ERROR:ERROR:  current transaction is aborted, commands ignored until end of transaction block
 [select nextid from ids where nodeid=1 and table_name='events' and field_name='eventid']
  6942:20101103:120301.160 [Z3005] Query failed: [0] PGRES_FATAL_ERROR:ERROR:  current transaction is aborted, commands ignored until end of transaction block
 [select max(eventid) from events where eventid between 100000000000000 and 199999999999999]
  6942:20101103:120301.160 [Z3005] Query failed: [0] PGRES_FATAL_ERROR:ERROR:  current transaction is aborted, commands ignored until end of transaction block
 [insert into ids (nodeid,table_name,field_name,nextid) values (1,'events','eventid',100000000000000)]
  6942:20101103:120301.162 [Z3005] Query failed: [0] PGRES_FATAL_ERROR:ERROR:  current transaction is aborted, commands ignored until end of transaction block
 [update ids set nextid=nextid+1 where nodeid=1 and table_name='events' and field_name='eventid']

What I changed recently (less than 24 hours ago) was setting in zabbix_server.conf :

MaxHousekeeperDelete=0

because of an enormous backlog in housekeeping (shutting down the zabbix_server took several hours!) based on Slow history housekeeping process on postgresql.

In the past I started zabbix 1.8.2 on postgresql 8.1 but migrated the database to 8.4.5 because of earlier performance issues.

What now? I'd like my monitoring back...

Update: a shutdown shows a large backlog again:

  6906:20101103:141902.503 Syncing history data...5.900863%
  6906:20101103:141915.626 Syncing history data...6.222729%
  6906:20101103:141927.725 Syncing history data...6.437305%

Update 2: using the reindex suggestions from the thread upgrade from 1.8.2 ->1.8.3 history updates way too slow and starting zabbix again gives me monitoring again. Let's see how this develops over the next days.

It still feels like there are 'costly' postgres queries happening.