Comments for Travis Hegner | Ramblings of an I.T. Guy

Comment on PostgreSQL 8.3 Warm Stand-by Replication by Peggy

Peggy — Tue, 24 Jan 2012 00:09:28 +0000

I have been trying to set this up in an 8.4 environment. When I startup the standby server and look at the standby.log, the Waiting for WAL file lists the WAL file that is the next in sequence for the standby server rather than the next log in sequence from the production server. How do I get it to look for the right WAL file. The %f parameter is providing the file name for the standby server. Has anyone else encountered this problem?

Comment on PostgreSQL 8.3 Warm Stand-by Replication by Abraão

Abraão — Tue, 17 May 2011 21:18:40 +0000

Nice work. I would like to know if you have some tips about warm replication in the Windows using the Postgres 8.2.11-1. Some material about it is very hard of find. Thanks very much.

Comment on PostgreSQL 8.3 Warm Stand-by Replication by Enzo

Enzo — Mon, 09 May 2011 16:42:09 +0000

“Unfortunately, there is no way to verify that the data is actually being replicated and recovered into the stand-by server, because during “recovery mode” the database is not operational.”
You can tail the pg_standby.log file and it will show the files that have been applied. Unfortunately though, this is pretty crappy in a production environment because you need to monitor this constantly and alert if you are too far behind. If you have a lot of databases, you aren’t going to be sitting around tailing log files. There needs to be a better way to monitor this and alert if you are too many log files behind.

Comment on Stevi & Trevor 3/5/11 The Best Man Toast by patti ann

patti ann — Mon, 07 Mar 2011 21:36:09 +0000

hey u did a great job and u made me cried to….. i love ya

Comment on Stevi & Trevor 3/5/11 The Best Man Toast by Kacey

Kacey — Mon, 07 Mar 2011 01:28:56 +0000

Travis, that seriously had me in tears. Absolutely beautiful!!

Comment on PostgreSQL 8.3 Warm Stand-by Replication by Chris Row

Chris Row — Sat, 25 Sep 2010 02:06:45 +0000

I was wondering if you ever got around to doing that write up on Slony. I’ve been working on implementing it myself on Debian systems, and while I can get it to mostly work I keep running into little snafus. Some due to the way we have to purge old data from the master database that would break Slony, and some on general setup.
I’m going over the docs again now but though if you had that write up it might save me some trouble. If I can’t figured it out I may have to fall back to WAL shipping for safety sake.
Thanks
-Chris

Comment on Adding Hbase 0.20 to Our Hadoop Cluster by Rizwan

Rizwan — Thu, 02 Sep 2010 13:56:50 +0000

hi,
How you will install hadoop on ubuntu enterprise cloud, well the reason i am asking this because i want to run to some experiments on multiple instances in cloud.
regards

Comment on PostgreSQL 8.3 Warm Stand-by Replication by Travis Hegner

Travis Hegner — Tue, 08 Jun 2010 19:17:56 +0000

Thanks Ian!
I assume that in case of a busy server, it keeps the last couple of WAL logs laying around in case they haven’t been read in yet. If the number of writes hits 16MB before your time limit, it will send an extra WAL file to the warm standby.
I have since moved from this method of replication to one based on Slony. I have been very pleased so far. Slony is a little bit harder to set up and maintain, but the fail-over is much smoother, and you can run backups/queries/reports from your slave, without impacting the master’s performance. I intend at some point to do a write up on it, but time is a scarce resource these days.
Thanks for reading!
Travis

Comment on PostgreSQL 8.3 Warm Stand-by Replication by Ian Douglas

Ian Douglas — Thu, 03 Jun 2010 00:14:55 +0000

Thanks for a great write-up. With a few small tweaks, I’ve successfully set up a replication node using your notes. Thanks for a very thorough write-up.
A few notes:
- Debian 5.0 (Lenny) needs “postgresql-contrib-8.3″ installed before the symlink from pg_wrapper to pg_standby will work.
- PG v8.3 renames the recovery.conf symlink file to recovery.done
- We’re seeing old logs hanging around for a two 10-minute cycles. When a new WAL shows up, we see console output like this:
running restore : OK
removing “/pg_import/000000010000000000000022″
Trigger file : /pg_import/role.master
Waiting for WAL file : 000000010000000000000025
WAL file path : /pg_import/000000010000000000000025
Restoring to… : pg_xlog/RECOVERYXLOG
Sleep interval : 60 seconds
Max wait interval : 0 forever
Command for restore : ln -s -f “/pg_import/000000010000000000000025″ “pg_xlog/RECOVERYXLOG”
Keep archive history : 000000010000000000000024 and later
… so it removed file 22 when 24 showed up, and updates its trigger to look for file 25, but file 23 and 24 hang around. I was concerned that I hadn’t done something correctly, but all appears well.

Comment on PostgreSQL 8.3 Warm Stand-by Replication by Travis Hegner

Travis Hegner — Wed, 14 Apr 2010 14:26:48 +0000

Thanks for all of the good feedback!
Be on the lookout as I have been exploring Slony-I a little, and I plan to write a tutorial on how to set that up.
Thanks,
Travis