Re: IDE Disk takes much IO wait

Martin Hierholzer <bespam.NO@xxxxxxxxxxx> wrote:
Here is some output which might help:

# hdparm -t /dev/sda
/dev/sda:
Timing buffered disk reads: 172 MB in 3.02 seconds = 57.04 MB/sec

I see no problem in your output. Whats wrong with 57 MB/s? Also, /dev/sda
doesn't sound as an IDE disk to me. I would guess that it is a SCSI disk
or SATA disk.

# lspci
....

So you have VIA chipset, a bt878 framegrabber and an nVidia graphics card.
Has any of this any relevance to your "problem"?


# cat /proc/interrupts
....

Nor do I understand how the number of interrupts or how different devices
get different interrupts relates to the "problem".

I have configured many other systems much in the same way (but with
other chipsets/CPUs), I never got those problems. I really don't have
ideas where to look anymore...

Before getting help to look for causes of the problem you must know what
the problem is. You claim that you have bad performance, but 57 MB/s
doesn't look like bad performance to me. Do you have any other benchmark
or example of a task that takes longer time than you think it should?

regards Henrik
--
The address in the header is only to prevent spam. My real address is:
hc1(at)poolhem.se Examples of addresses which go to spammers:
root@localhost postmaster@localhost

.

Relevant Pages

  • Re: IDE Disk takes much IO wait
    ... He never said that 57 MB/s was the problem. ... doesn't sound as an IDE disk to me. ... Then you're not familiar with how some disk controller drivers act. ... Nor do I understand how the number of interrupts or how different devices ...
    (comp.os.linux.hardware)
  • Re: 10g RAC: max performance & min cost with miSCSI?
    ... :-) Due to reliability and minimun downtime requirements we are planning to build a 2 node Linux Oracle 10g RAC. ... To minimize the cost we are planning to run the nodes 1 CPU each since Oracle lisences are per CPU. ... The bottleneck seems almost always be on the disk system so starting the design on disk system seems the right way to go. ... Since we have 2 CPUs on RAC the disk system should be able to deliver 400 MB/s, let's say 500 MB/s to be safe. ...
    (comp.databases.oracle.server)
  • 10g RAC: max performance & min cost with miSCSI?
    ... To minimize the cost we are planning to run the nodes 1 CPU each since Oracle lisences are per CPU. ... The bottleneck seems almost always be on the disk system so starting the design on disk system seems the right way to go. ... Since we have 2 CPUs on RAC the disk system should be able to deliver 400 MB/s, let's say 500 MB/s to be safe. ...
    (comp.databases.oracle.server)
  • Re: Help me, a friend of mine wrote a program in C# and I wrote the same program in C..His is faster
    ... DiskIO 300.55 Mb/s ... fwritereturns after putting the data into a buffer but before the hardware actually acknowledges it's been written to disk. ... If an OS don't flush system buffers when a file is closed, or is doing a fflushin an async manner, then it is impossible to code a bullet-proof C program with file I/O, in that environment. ... It's not so amazing when you realize you're measuring the speed of the OS's I/O system, ...
    (comp.lang.c)
  • Re: SATA-performance: Linux vs. FreeBSD
    ... I used a dedicated disk for this task. ... I have a flash disk from a manufacturer who grants me 48 MB/s. ... or alternate between discs when writing. ... Real Programmers consider "what you see is what you get" to be just as ...
    (Linux-Kernel)