Re: I/O - Increase RAM - Mailing list pgsql-general

From FarjadFarid\(ChkNet\)
Subject Re: I/O - Increase RAM
Date
Msg-id 000901d195d8$26b6be00$74243a00$@checknetworks.com
Whole thread Raw
In response to I/O - Increase RAM  ("drum.lucas@gmail.com" <drum.lucas@gmail.com>)
List pgsql-general

Sorry, by the first solution I meant to separate the drive for your db files from other tasks.

 

/////////////

 

It sounds as if you have *one* server with 128GB of memory.

 

Therefore the probabilities are that you have too many tasks running at the same time. Flushing the disk drives’ read/write cache.

 

Increasing the memory probably will not solve the problem. The most effective solution would be to separate the tasks that access the drive (read or write) often.

 

Alternatively if you have access to the code. Try to cache some of the disk activities.  

 

The third solution would be to separate the tasks and move them to a different server.

 

Hope this helps.

 

 

 

 

From: pgsql-general-owner@postgresql.org [mailto:pgsql-general-owner@postgresql.org] On Behalf Of drum.lucas@gmail.com
Sent: 13 April 2016 21:44
To: Postgres General
Subject: [GENERAL] I/O - Increase RAM

 

Hi all,

 

At the moment I'm having 100% I/O during the day. My server has SATA HDs, and it can't be changed now.

So, to solve the problem (or at least try) I was thinking about double the RAM, and by doing that, increasing the cache.

 

Inline images 1

 

The server has 128GB today:

 

shared_buffers = 51605MB (I'll change it to 32GB)
effective_cache_size = 96760MB

 

 

Question:

 

I know that might not be the best option, but by increasing the RAM and the CACHE would help, right?

 

Thanks

Lucas

Attachment

pgsql-general by date:

Previous
From: Kevin Grittner
Date:
Subject: Re: pgpool-II: cannot use serializable mode in a hot standby
Next
From: Tatsuo Ishii
Date:
Subject: Re: pgpool-II: cannot use serializable mode in a hot standby