Hi
If I understood you the right way, you can't understand, why the linux
box is much faster than the NT? If I am wrong, stop reading, if I am
right proceed.
1) We also have an Oracle instance running on a linux box (nearly same
configuration, except we have two raid0 - four IBM u2W disks each - on
two Adaptec U2W), and benchmarked its performance in comparison to our
HP (K, D and C) machines and our Nile (Pyramid Technologies).
The linux box is 20 times faster than the Nile (cluster of two Niles,
ten 180Mhz CPUs each) concerning I/O performance and roundabout 5 to 10
times faster concerning queries and load.
In comparison to a double cpu HP C it is as fast as the HP machine
concerning queries and 2 times faster concerning I/O performance (the HP
got same controllers, and two 360 MHz CPUs).
2) The output of Bonnie is very difficult to understand, and I am of the
opinion, you must have detailed kernel (fs) knowledge to understand it
the right way; hdparm seems to be easier and more "realistic". An I/O
performance of 12 MB/s on raid5 is very good, I think. The raid
controller (in oposite of a software raid) has to calculate the checksum
of the data and to serve the disk I/O at the same time. We made
benchmarks with Bonnie on our stripes (raid0) and got round about 15
MB/s with 2GB-files; at first I wondered about the "low" performance,
but after doing the benchmarks against HP, I know, what 15 MB/s means.
3) I got two workstations at my working place: one is a PPro200 (32MB
PS/2 RAM) with an three or four year old IDE disk, the other is a
PII-266 (128 MB SD-RAM) with an one year old UW-SCSI disk. The Pro is a
linux box and the PII a NT4.0 box; booting the workstations at the same
time, results in a wait for the NT. Starting Netscape on both at the
same time results in a wait for the NT. Copying a text file of same size
on both machines from one directory to the other results in - guess it.
To sum up: I think linux is able to get as much performance out of the
components as possible (for which it is known); NT does not.
Greetings, Dietmar
p.s: do you got the Oracle Intelligent Agent running on your box? We had
problems to start, but now it seems to work. Any comments towards
installation of the Agent?
"Soukup, Kevin" wrote:
>
> Hello linux-raid!
>
> I'm a newbie, so please forgive any stupidity. I've searched and read
> diligently and feel I need to ask this more specifically. I have:
>
> Dell 6300
> Dual 400 mHz Xeon w/512K cache each
> 1GB RAM
> Adaptec SCSI controller
> Dell PERC 16MB (AMI MegaRAID)
> Redhat 6.0 vanilla (2.2.5-15)
> 6-18GB Quantum 7200 RPM in
> RAID 5
>
> I've been benchmarking Oracle on this server against an identical
> Oracle instance on an IDENTICALLY EQUIPPED NT server. The Linux box
> KICKS the NT's butt, drastically, on all but the simplest, least
> loaded queries. I've got statistics if anyone is interested.
>
> Given the Linux box is so much faster than the NT server, and given
> I'm using a (reputedly) slow RAID controller, I'd like to verify
> something before proceeding. IF this RAID controller is so slow,
> I can imagine the performance margin would increase if I could
> increase the RAID performance.
>
> All the reading I've been doing would suggest that the MagaRAID
> controller is the slowest RAID solution I could choose. Based on
> the Bonnie stats below, I THINK I recognize what is meant. The
> final two Bonnie runs on files of size 2047GB look like the IO is
> a serious bottleneck.
>
> But I need help and validation to continue.
>
> 1) Am I correct in reading the Bonnie numbers to say the greatest
> throughput (using a file of size 2047GB) is 9.655MB during the
> Sequential Block Input, and does that mean "write" or "read"?
>
> 2) Regarding the hdparm output, is it saying the best disk read
> is 12.08 MB/sec?
>
> 3) And does this reinforce the Bonnie stats?
>
> 4) Isn't this all pretty slow?
>
> 5) if so, is there anything I can do about it?
>
> -- ===================================================================== --
> -- I ran Bonnie against the described server: --
>
> --------Sequential Output-------- ---Sequential Input---- ---Random---
> -Per Char- ---Block--- -Rewrite-- -Per Char- ---Block---- ---Seeks----
> MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
> 128 4793 77.7 62438 100.0 12211 41.8 6177 98.3 174710 100.0 8344.5 200.3
> 128 6112 99.0 62926 99.4 31905 96.9 6205 98.3 154055 97.6 8362.8 200.7
> 256 4097 66.1 30336 47.0 18749 68.2 6197 98.5 161505 99.8 7746.3 199.5
> 256 6081 99.3 36635 55.8 16771 57.4 6185 97.9 161112 100.8 7698.8 190.5
> 512 3983 65.4 13016 21.7 7956 36.8 6214 99.0 140129 100.2 2839.2 75.9
> 512 6026 97.9 12731 22.2 7931 32.6 6239 99.0 140255 100.1 2821.9 75.5
> 1024 4796 77.8 8991 17.2 4615 18.5 5103 80.6 10319 19.0 408.7 10.4
> 1024 5646 92.4 9094 17.9 5081 21.5 4911 77.5 8887 18.2 396.5 9.5
> 2047 4630 75.7 8577 18.6 4559 18.0 5274 82.6 9655 20.1 98.7 3.9
> -- ===================================================================== --
> -- I ran hdparm and got these results: --
>
> [root@patty system_stats]# hdparm -tT /dev/sda5 /dev/sda6
>
> /dev/sda5:
> Timing buffer-cache reads: 64 MB in 0.62 seconds =103.23 MB/sec
> Timing buffered disk reads: 32 MB in 2.65 seconds = 12.08 MB/sec
>
> /dev/sda6:
> Timing buffer-cache reads: 64 MB in 0.64 seconds =100.00 MB/sec
> Timing buffered disk reads: 32 MB in 2.69 seconds = 11.90 MB/sec
> -- ===================================================================== --
>
> Any and all input will be greatly appreciated. You may respond directly if you
> don't want to bother the rest of the list with this (probably) infantile question.
>
> Best regards,
>
> Kevin
>
> =========================================
> Kevin Soukup
> phone: (815) 235-6944 (wk)
> phone: (815) 232-4133 (hm)
> FAX: (815) 235-5959 (wk)
> e-mail: [EMAIL PROTECTED]
> =========================================
