pg.ddx.io  pgsql-admin@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: brian <brian@teamadr.com>
To: pgsql-admin@postgresql.org
Subject: Re: New system recommendations
Date: Thu, 27 Apr 2006 15:11:30 -0700
Message-ID: <44514192.2010909@teamadr.com> (raw)
In-Reply-To: <20060427050605.GW97354@pervasive.com>
References: <8511B4970E0D124898E973DF496F9B432515CF@stash.stackdump.local>
	<20060427050605.GW97354@pervasive.com>

Jim C. Nasby wrote:
> On Wed, Apr 26, 2006 at 07:30:49PM -0600, Benjamin Krajmalnik wrote:
>   
>> I am about to take a system from testing into production.
>>
>> The system is a combination oltp/bi (network monitoring platform).
>>
>> We are currently inserting about 1 million rows per day, and will
>> increase to probably 5 million once it goes into full deployment.
>>
>> The current test production server is running Windows Server 2003,
>> PostgreSQL 8.1.3.
>>
>> Record insertion is done via an ODBC call to a stored procedure from a
>> Windows based host monitor.
>>     
>
> Make sure you're batching inserts with transactions.
>
>   
>> It is currently running on a 3GHz Xeon HT, 2GB RAM, dual 72GB disks
>> running RAID 1.  This server is a 1U without only 2 drive bays, so I
>> have a potential issue with drive space.
>>
>> As a result, I will be moving the db server to a Dell 1650 with 3 146GB
>> SCSI drives running RAID 0.  System is a dual processor, 1.2GHz, with
>> 4GB RAM.
>>     
>
> 3 drive raid0 is likely to fail within 3-4 years, just so you know;
> unless it's not new hardware, in which case I'd expect something closer
> to 2 years (my general experience is that server HDs will last 4-6
> years, so with 3 of them you're looking at a failure every ~2 years).
> You sure you want to trust a monitoring app to raid0? :)
>
>   
>> Due to the amount of record insertions being performed, record insertion
>> speed is paramount - also because excessive execution time will have a
>> side effect of causing the monitoring agent to go stale.  Some of the
>> users have mentioned that I will get better performance running under a
>> *nix OS.  We are mostly a Microsoft OS house, but also run FreeBSD.  I
>> am considering deploying with FreeBSD 6.0.  I was wondering if anyone
>> has benchmarks showing speed of execution of PostgreSQL 8.1.3 under
>> Win2003 and FreeBSD 6.0.  Also, are there any caveats or items I should
>> be aware of if running under FreeBSD?  Any issues when running under a
>> multi-processor kernel?  Anything in specific which I should include in
>> the kernel build to give me optimum performance for running PostgreSQL?
>>
>> Needless to say, I am a bit nervous of moving to FreeBSD since I have
>> not tested it in a production environment.
>>     
>
> Aside from some very large shops running FreeBSD (yahoo does I believe,
> and hotmail did for a long time, even after MS bought it), it's also a
> preferred OS by PostgreSQL developers. On the other hand we've only had
> Windows support for about 2 years and there's known windows-only issues.
>   
Not certain about freebsd 6.x, but on 4.x, running an intel SMP build
limits each machine to three and a half gigs of memory:

FreeBSD 5.4-STABLE #0: Fri Sep  2 11:32:58 PDT 2005
Timecounter "i8254" frequency 1193182 Hz quality 0
CPU: Intel(R) Xeon(TM) CPU 3.06GHz (3056.50-MHz 686-class CPU)
  Origin = "GenuineIntel"  Id = 0xf27  Stepping = 7

Features=0xbfebfbff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CLFLUSH,DTS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,PBE>
  Hyperthreading: 2 logical CPUs
real memory  = 3759144960 (3585 MB)
avail memory = 3679334400 (3508 MB)

There is 4 gigs of physical memory in the machine that produced the
above.  PAE cannot be built into this kernel... I can't even remember
what the error was..
I think PAE ended up being incompatible with HTT.

-Brian




view thread (15+ messages)  latest in thread

Message-ID: <44514192.2010909@teamadr.com>
Permalink:  ../44514192.2010909@teamadr.com/
Also on:    postgresql.org/message-id/44514192.2010909@teamadr.com

 · 

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients from your mail client; the link fills in
  the headers below: reply via email
  Some clients, webmail especially, drop In-Reply-To and References from
  a link, and the reply then starts a new thread: use the mbox below.

  To: pgsql-admin@postgresql.org
  Cc: brian@teamadr.com
  Subject: Re: New system recommendations
  In-Reply-To: <44514192.2010909@teamadr.com>
  References: <8511B4970E0D124898E973DF496F9B432515CF@stash.stackdump.local> <20060427050605.GW97354@pervasive.com> <44514192.2010909@teamadr.com>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

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