pg.ddx.io  pgsql-admin@postgresql.org mailing list archive  
help / color / mirror / Atom feed
Stripping raid vs. table spaces
3+ messages / 3 participants
[nested] [flat]

* Stripping raid vs. table spaces
@ 2005-06-09 07:16 KÖPFERL Robert <robert.koepferl@sonorys.at>
  2005-06-09 11:14 ` Re: Stripping raid vs. table spaces Dawid Kuroczko <qnex42@gmail.com>
  2005-06-09 14:11 ` Re: Stripping raid vs. table spaces Scott Marlowe <smarlowe@g2switchworks.com>
  0 siblings, 2 replies; 3+ messages in thread

From: KÖPFERL Robert @ 2005-06-09 07:16 UTC (permalink / raw)
  To: pgsql-admin

Short question:

Is it enugh to have stripping on a disk array?
Are table spaces then of less use than the actual stripping? So will just
stripping give me more O/I-performance?



^ permalink  raw  reply  [nested|flat] 3+ messages in thread

* Re: Stripping raid vs. table spaces
  2005-06-09 07:16 Stripping raid vs. table spaces KÖPFERL Robert <robert.koepferl@sonorys.at>
@ 2005-06-09 11:14 ` Dawid Kuroczko <qnex42@gmail.com>
  1 sibling, 0 replies; 3+ messages in thread

From: Dawid Kuroczko @ 2005-06-09 11:14 UTC (permalink / raw)
  To: KÖPFERL Robert <robert.koepferl@sonorys.at>; +Cc: pgsql-admin

On 6/9/05, KÖPFERL Robert <robert.koepferl@sonorys.at> wrote:
> Short question:
> 
> Is it enugh to have stripping on a disk array?
> Are table spaces then of less use than the actual stripping? So will just
> stripping give me more O/I-performance?

Short answer:  if you make one striping partition -- you have better
overall performance.

if you distribute tables using tablespaces -- you make tables
performance more independant from other tables.  I.e. large
load on one table has less impact on the load on the other
table.

Personally I would keep indexes on one set of drives and data
on other.  All depends on your setup.



^ permalink  raw  reply  [nested|flat] 3+ messages in thread

* Re: Stripping raid vs. table spaces
  2005-06-09 07:16 Stripping raid vs. table spaces KÖPFERL Robert <robert.koepferl@sonorys.at>
@ 2005-06-09 14:11 ` Scott Marlowe <smarlowe@g2switchworks.com>
  1 sibling, 0 replies; 3+ messages in thread

From: Scott Marlowe @ 2005-06-09 14:11 UTC (permalink / raw)
  To: KÖPFERL Robert <robert.koepferl@sonorys.at>; +Cc: pgsql-admin

On Thu, 2005-06-09 at 02:16, KÖPFERL Robert wrote:
> Short question:
> 
> Is it enugh to have stripping on a disk array?
> Are table spaces then of less use than the actual stripping? So will just
> stripping give me more O/I-performance?

The answer is, it depends.

If you had some large number of drives to put in a RAID array, along
with a fast RAID controller and lots of battery backed cache set to
write-back, then it is unlikely you would get any better performance
from putting tables and indexes on lots of different discrete drives. 
Of course, you can use both, putting table spaces on top of different
stripe sets.

The important thing to consider is your usage pattern.  Will you be
writing in several small streams and reading in one or two big ones?  Or
vice versa?  Or will you have many read/write transactions in transit at
once?  Will you have a few heavy traffic tables and lot of low traffic
ones?  On and on.

There's a big difference between a data warehouse / OLAP and a
transactional engine / OLTP setup.

Note that there is probably better info on this in the -performance
archives than in the -admin ones...



^ permalink  raw  reply  [nested|flat] 3+ messages in thread


end of thread, other threads:[~2005-06-09 14:11 UTC | newest]

Thread overview: 3+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2005-06-09 07:16 Stripping raid vs. table spaces KÖPFERL Robert <robert.koepferl@sonorys.at>
2005-06-09 11:14 ` Dawid Kuroczko <qnex42@gmail.com>
2005-06-09 14:11 ` Scott Marlowe <smarlowe@g2switchworks.com>

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox