Hosting Provided By

Re: amcc/3ware

From: Boris Pavlov <edi(at)elib.minfin.bg>
Date: Fri Jan 11 2008 - 12:39:47 EST

Andrew Miehs wrote:
>
> On 11/01/2008, at 3:19 PM, Boris Pavlov wrote:

>>>>>
>>>>
>>>> I am not a big fan of 3ware Raid controllers. Up until now it was 
>>>> because
>>>> they were just plan SLOW...
>>
>> just a quick word from me - simple, can not agree that they are slow. 
>> they are not SLOW (caps on) either - comparing to soft raid - same 
>> raid level on same computer.

>
> I am glad that you have had better success with them than I have.
>
> My 'slow' experience comes from 2 Linux servers running
>
> [ 40.834489] 3ware Storage Controller device driver for Linux
> v1.26.02.001.
> [ 46.993597] scsi2 : 3ware Storage Controller
> [ 46.993692] 3w-xxxx: scsi2: Found a 3ware Storage Controller at
> 0xafa0, IRQ: 22.
> [ 46.993980] Vendor: 3ware Model: Logical Disk 0 Rev: 1.2
> [ 46.995418] Type: Direct-Access ANSI SCSI
> revision: 00
> [ 46.995952] Vendor: 3ware Model: Logical Disk 2 Rev: 1.2
> [ 46.997401] Type: Direct-Access ANSI SCSI
> revision: 00
> [ 46.999678] 3ware 9000 Storage Controller device driver for Linux
> v2.26.02.004.
>
> 0000:02:02.0 RAID bus controller: 3ware Inc 3ware 7000-series ATA-RAID
> (rev 01)
> Subsystem: 3ware Inc 3ware Inc 3ware 7xxx/8xxx-series PATA/SATA-RAID
> Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 22
> I/O ports at afa0 [size=16]
> Memory at fe7ffff0 (32-bit, non-prefetchable) [size=16]
> Memory at fd800000 (32-bit, non-prefetchable) [size=8M]
> Expansion ROM at fe7e0000 [disabled] [size=64K]
> Capabilities: [40] Power Management version 1
>
> These cards are about 4 years old now - and the disks directly connected
> to the
> SATA controller are quicker than the the same disks connected via the 3ware
> controller - timing with 'dd' and hdparm.
>

two questions -

1.) did you read the suggestions on the 3ware's site? 2.) and second, sorry i do not want to be abusive, but how long (as uptime) after the creation of the volume you actually did test the controller?

did you check the status of the array with tw_cli?

almost as i tough - sometimes the check show less performance, but in general the software raid performs really bad on a burdened server - i've seen this and actually was swearing when, say, about 5 years ago purchased my first 3ware (pata pci) board, which actually solved my problems. afaicr, the board was about 10-15% slower than hdparm with the disks attached to the main board ata. but this is not the same case when the load goes high, believe me :) the server used to nearly hang before that.

hdparm is not quite relevant, by the way...

edi
PS if you want, we can discuss the details in private - i do use several different amcc/3ware boards, and any possible problems are catching my attention...

> And I have had similar experience on 1 windows server - which is running
> the same card.
> - Windows 2003 running as a file server.
> - This will be replaced in the next 3 months with a NetAPP
>

Do you need help?

[cut]

> Personally - I would use some sort of SAS/ SCSI Raid for the boot disks,
> and use an external storage such
> as 'Infortrend' or NetApp depending on budget. I really dislike internal
> mass storage.
>
> Cheers
>
> Andrew
>
>

-- 
To UNSUBSCRIBE, email to debian-isp-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

Received on Fri Jan 11 12:40:22 2008

This message: [ Message body ]
Next message: Seth Mattinen: "Re: 3ware Raid controllers and 64bit Etch"
Previous message: Andrew Miehs: "Re: New Mailserver..."
In reply to: Andrew Miehs: "Re: New Mailserver..."
Next in thread: Andrew Miehs: "Re: amcc/3ware"
Reply: Andrew Miehs: "Re: amcc/3ware"

Contemporary messages sorted: [ By Date ] [ By Thread ] [ By Subject ] [ By Author ] [ By messages with attachments ]

This archive was generated by hypermail 2.1.8 : Wed Mar 19 2008 - 06:52:11 EDT