[Spacewalk-list] Migration from pgsql to oracle

[Boyd, Robert]

Paul Robert,

This sounds on the face of it like an interesting possibility.   On the other hand, I looked at the page you linked to.  There I see many technical points that require an understanding of database configuration and administration that I certainly don't possess at this point in time.   I think it's asking a bit much of a non-DBA to come up to speed to tackle such a task to get better performance out of Spacewalk.

It would be more beneficial for someone who has already done partitioning to offer a description of the steps they went through including any calculations they had to make with their own Spacewalk database so that someone else could replicate the process.

Robert Boyd
Sr. Systems Engineer 
PeopleFluent
p. 919-645-2972 | c. 919-306-4681
e. Robert.Boyd at PeopleFluent.com


Click here to experience the power of the new PeopleFluent Mirror Suite ™

Visit: www.peoplefluent.com | Read: PeopleFluent Blog | Follow: @PeopleFluent

-----Original Message-----
From: spacewalk-list-bounces at redhat.com [mailto:spacewalk-list-bounces at redhat.com] On Behalf Of Paul Robert Marino
Sent: Wednesday, November 19, 2014 10:51 AM
To: spacewalk-list at redhat.com
Subject: Re: [Spacewalk-list] Migration from pgsql to oracle

Have you considered using a less drastic approach like partitioning some of the tables and tuning the ram.
I some how doubt the single CPU core per query it the problem because PostgreSQL is very good at handling bulk queries as opposed to lots of queries for tiny data sets.

Here is something you can check if  a directory named /var/lib/pgsql/data/pg_tmp exist on your system then your queries are exceeding the allowed rem for a single query and PostgreSQL is using that directory to essentially create swap files for queries if that is happening you need more ram or you need to tune your PostgreSQL server.

Here is a good place to start reading about table partitions https://wiki.postgresql.org/wiki/Table_partitioning that with constraint exclusion enabled in the config can help a lot with large data sets.