Re: 8x2.5" or 6x3.5" disks - Mailing list pgsql-performance

From Mike Smith
Subject Re: 8x2.5" or 6x3.5" disks
Date
Msg-id 51494DB187D98F4C88DBEBF1F5F6D42303EC56C9@edb06.mail01.enterprisedb.com
Whole thread Raw
In response to 8x2.5" or 6x3.5" disks  ("Christian Nicolaisen" <blackbrrd@gmail.com>)
Responses Re: 8x2.5" or 6x3.5" disks
Re: 8x2.5" or 6x3.5" disks
List pgsql-performance

You don’t mention the capacity of the disks you are looking at. Here is something you might want to consider.

 

I’ve seen a few performance posts on using different hardware technologies to gain improvements. Most of those comments are on raid,  interface and rotation speed.   One area that doesn’t seem to have  been mentioned  is to  run your disks empty.

 

One of the key roadblocks in disk performance is the time for the disk heads to seek, settle and find the start of the data. Another is the time to transfer from disk to interface.  Everyone may instinctively know this but its often ignored.

 

Hard disks are CRV ( constant rotational velocity) = they spin at the same speed all the time

Hard disk  drives use a technology called ZBR  = Zone Bit Recording =  a lot more data on the outside tracks than the inner ones.

Hard disk fill up from outside track to inside track  generally unless you’ve done some weird partitioning.   

 

On the outside of the disk you get a lot more data per seek than on the inside. Double whammy you get it faster.

Performance  can vary more than  100% between the outer and inner tracks of the disk.   So running a slower disk twice as big may give you more benefit than running a small capacity 15K disk full.  The slower disks are also generally more reliable and mostly much cheaper.

 

The other issue for full disks especially with lots of random small transactions is the heads are seeking and settling  across the whole disk but  typically with most of those seeks being on the latest transactions which are placed nicely towards the middle of the disk.

 

I know of a major bank that has a  rule of thumb 25% of the disk partioned  as  a target maximum for high performance disk systems in a key application. They also only pay for used capacity from their disk vendor.   

 

This is not very green as you need to buy more disks for the same amount of data and its liable to upset your purchasing department who won’t understand why you don’t want to fill your disks up.

 

Mike

 

pgsql-performance by date:

Previous
From: Arjen van der Meijden
Date:
Subject: Re: 8x2.5" or 6x3.5" disks
Next
From: Euler Taveira de Oliveira
Date:
Subject: Re: Performance problems inside a stored procedure.