BUG #8534: Missing record in binary replica 9.3

From: kavehmz(at)gmail(dot)com
To: pgsql-bugs(at)postgresql(dot)org
Subject: BUG #8534: Missing record in binary replica 9.3
Date: 2013-10-17 15:48:16
Message-ID: E1VWpo8-0008LK-Lx@wrigleys.postgresql.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

The following bug has been logged on the website:

Bug reference: 8534
Logged by: Kaveh Mousavi Zamani
Email address: kavehmz(at)gmail(dot)com
PostgreSQL version: 9.3.0
Operating system: Debian 6.0
Description:

All,

I found something in the following setup:

I have 3 DBS

db01 (9.3.0) master
db02 (9.3.0) replica same network
db03 (9.3.0) sync replica another network 80ms away.

db01 have around 30 - 300 records per second update.

At some point today I found there is a missing record in both replica.
Data after that point has been replicated though

db01
select * from transaction.account where client_detail_id ='163042';
id | client_detail_id | last_modified
---------+----------------+---------------+----------+
1736861 | 163042 | 2013-10-17 10:39:18.312815

in both other dbs no record.

Does anyone has any idea how could this happen? Is this a bug in 9.3
replication?

then in db01 I did update last_modified field for that record and suddenly
new record and new last_modified values appeared in both replicas.

I double check the replica DB and I can confirm there are more missing
lines in different tables.

Regards

Browse pgsql-bugs by date

  From Date Subject
Next Message digoal 2013-10-17 23:19:13 Re: BUG #8531: systemtap probe mark(checkpoint__done) error when i read the parameters
Previous Message Alvaro Herrera 2013-10-17 15:47:54 Re: BUG #8531: systemtap probe mark(checkpoint__done) error when i read the parameters