Re: Harddrive performance: Promise SATA RAID under Redhat 9
From: emilio lazardo (NotARealAddress_at_spam.this)
Date: 12/12/03
- Next message: mjt: "Re: File size exceeded"
- Previous message: Baho Utot: "Re: File size exceeded"
- In reply to: Ken Kauffman: "Re: Harddrive performance: Promise SATA RAID under Redhat 9"
- Next in thread: Alon Solell: "Re: Harddrive performance: Promise SATA RAID under Redhat 9"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Thu, 11 Dec 2003 23:48:55 GMT
>>"Ken Kauffman" <kkauffman@nospam.headfog.com> wrote in message
>>news:eB9Bb.46823$dO2.19889@lakeread03...
>> ...
> all depends on your application. But being on 24 hour call for a data
> center, I will not put IDE in a critical system. Also, does 3Ware have a
> hot swappable/rebuildable solution for IDE?
3Ware controllers have hot spare and automatic rebuild. Hot swappable also
depends on the PCI bus/motherboard/driver combination, not just the controller.
I've used a 3Ware 6400 (ATA/66) in a small server for several years: 2.2.x
and 2.4.x kernels (which include the 3W driver) as an individual drive
controller, stripe and mirror sets, and a RAID 5 with spare configuration.
I've had zero problems with the controller in any configuration.
As for 3w performance, striping yields near 2x overall performance with the
usual danger of losing a disk. Mirroring works as expected except that the
mirror pairs are not portable to other controllers. RAID 5 was a tad slow.
In general a 10-15% performance increase over other controllers. For
example, Promise U/66 controller on the identical setup is 15% slower,
requires 3x CPU and tends to lock up under load.
> I'm not trying to be argumentative, I just am weighing the offering at an
> enterprise class level.
If you really want enterprise class go with offerings from Netapp, IBM or
similar but be willing to pay.
rgds,
tim.
3W + ancient KA7 + .18u Athlon @900MHz, software RAID 0, 2.4.23 kernel:
[tim@abit tim]# time hdparm -tT /dev/md0
/dev/md0:
Timing buffer-cache reads: 664 MB in 2.01 seconds = 330.35 MB/sec
Timing buffered disk reads: 206 MB in 3.01 seconds = 68.44 MB/sec
0.260u 3.430s 0:15.19 24.2% 0+0k 0+0io 1130pf+0w
...
PCI: Found IRQ 15 for device 00:0d.0
scsi1 : Found a 3ware Storage Controller at 0xe400, IRQ: 15, P-chip: 5.7
scsi1 : 3ware Storage Controller
Vendor: 3ware Model: Logical Disk 0 Rev: 1.0
Type: Direct-Access ANSI SCSI revision: 00
Vendor: 3ware Model: Logical Disk 1 Rev: 1.0
Type: Direct-Access ANSI SCSI revision: 00
Attached scsi disk sda at scsi1, channel 0, id 0, lun 0
Attached scsi disk sdb at scsi1, channel 0, id 1, lun 0
SCSI device sda: 40188960 512-byte hdwr sectors (20577 MB)
sda: sda1 sda2 sda3 < sda5 sda6 sda7 sda8 >
SCSI device sdb: 40188960 512-byte hdwr sectors (20577 MB)
sdb: sdb1 sdb2 sdb3 < sdb5 sdb6 sdb7 sdb8 >
...
[tim@abit tim]# cat /proc/mdstat
Personalities : [raid0]
read_ahead 1024 sectors
md0 : active raid0 sdb5[1] sda5[0]
4192768 blocks 32k chunks
md1 : active raid0 sdb6[1] sda6[0]
2425600 blocks 32k chunks
md2 : active raid0 sdb7[1] sda7[0]
12707200 blocks 64k chunks
md3 : active raid0 sdb8[1] sda8[0]
20113152 blocks 64k chunks
unused devices: <none>
(backed up with rsync snapshots)
- Next message: mjt: "Re: File size exceeded"
- Previous message: Baho Utot: "Re: File size exceeded"
- In reply to: Ken Kauffman: "Re: Harddrive performance: Promise SATA RAID under Redhat 9"
- Next in thread: Alon Solell: "Re: Harddrive performance: Promise SATA RAID under Redhat 9"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
