Skip site navigation (1) Skip section navigation (2)

testing cvs HEAD - HS/SR - xlog timeline 0 pg_xlogfile_name_offset

From: "Erik Rijkers" <er(at)xs4all(dot)nl>
To: pgsql-hackers(at)postgresql(dot)org
Subject: testing cvs HEAD - HS/SR - xlog timeline 0 pg_xlogfile_name_offset
Date: 2010-02-24 19:40:14
Message-ID: (view raw or flat)
Lists: pgsql-hackers
I have two 9.0devel machines (*not* alpha but cvs as of 2010.02.19 22.48).

One primary, one slave.

In an attempt to keep track of standby progression (versus primary), I ran slave-side:

replicas=# select
,                           pg_last_xlog_receive_location()
,   pg_xlogfile_name_offset(pg_last_xlog_receive_location())
 pg_last_xlog_replay_location |      pg_xlogfile_name_offset       |
 E2/C012AD90                  | (00000000000000E2000000C0,1224080) |

 pg_last_xlog_receive_location |      pg_xlogfile_name_offset
 E2/C012AD90                   | (00000000000000E2000000C0,1224080)
 (1 row)

These zero-timeline filenames look suspicious, no?
I understand timeline-count to normally start at 1, not 0?

The replication seems to be running fine (680 GB).

ps seems to report the right xlog filename (slave):

/var/data1/pg_stuff/pg_installations/pgsql.sr_hotslave/bin/postgres -D
 \_ postgres: startup process   recovering 00000001000000E2000000C0
 \_ postgres: wal receiver process   streaming E2/C012AE28
 \_ postgres: writer process
 \_ postgres: stats collector process

replicas=# select version();
 PostgreSQL 9.0devel on x86_64-unknown-linux-gnu, compiled by GCC gcc (GCC) 4.4.3, 64-bit
(1 row)

Is the filename that pg_xlogfile_name_offset( pg_last_xlog_(replay|receive)_location() ) reports a
bug, or expected as shown?


Erik Rijkers


pgsql-hackers by date

Next:From: RĂ©mi ZaraDate: 2010-02-24 19:45:12
Subject: Re: NaN/Inf fix for ECPG
Previous:From: Bruce MomjianDate: 2010-02-24 19:35:14
Subject: Re: dblink for 8.4 should work without user-mappings

Privacy Policy | About PostgreSQL
Copyright © 1996-2014 The PostgreSQL Global Development Group