Re: linux os cache and ocfs

Hi Mladen,

Thank you very much for your help.

I have logged a tar related to the direct io - ocfs issue to oracle . Tech. support engineer told me that since i use ocfs it use direct io no matter the filesystemio_options parameter. But you say the opposite.

>No, it doesn't. By telling oracle to use direct I/O, you are
> adding O_DIRECT to the arguments of the "open" system call.

Is it possible to do strace to one of the executables of oracle and try to see the o_direct flag?

Kind Regards,
tolga

Mladen Gogala <gogala_at_sbcglobal.net> wrote in message news:<pan.2004.07.29.15.29.27.147902_at_sbcglobal.net>...
> On Thu, 29 Jul 2004 04:35:13 -0700, utkanbir wrote:
>
> > Hi Mladen ,
> >
> > Thanks for your answer. In fact your answer resulted another
> > questions.
> >
> >> It uses direct I/O only if you activate it and only if it is
> >> supported in kernel. You can activate direct I/O by setting the
> >> parameter filesystemio_options appropriately.
> >
> > My setting for filesystemio_options is none .
>
> This is why direct I/O is not being used.
>
> > I was thinking to set it
> > as async in order to enable aio till you had told me about the direct
> > i/o. Whats the difference between direct i/o and async .
>
>
> Asynchronous I/O takes place in the background and the issuing process
> doesn't wait for the I/O to complete, but is notified instead. Direct I/O
> transfers data directly from the disk into the user buffer, bypassing the
> buffer cache.
>
>
> > Just ignoring
> > this parameter , if i use ocfs , doesnt it open files by using
> > o_direct?
>
> No, it doesn't. By telling oracle to use direct I/O, you are
> adding O_DIRECT to the arguments of the "open" system call.
>
> > And doesnt this mean that the io to the file does not use
> > any os buffer?
>
> No. It means that the data is transfered from the driver buffer (1MB in
> size) directly to the user buffer.
>
>
> >
> > In fact i mean paging not swapping. As far as i know , unix systems
> > does not use swapping but only paging . I really want to ask the
> > reason of swap disk space usage. Consider a new process , when it
> > first starts running , only the working set of this process must be in
> > physical memory ( so thats why its called working set ) .When pages
> > which are not in memory are needed , page fault occurs and data must
> > be transferred from the disk. When this occurs , does the swap usage
> > value in top display increase? Isnt it used in order to satisfy
> > virtual memory paging ? Isnt it pagefile?(or vmstat si so values).
> > In my case isnt my paging value high?
>
> Paging and swapping are two completely different phenomenons. Paging
> deals with pages, swapping deals with processes. Page stealing will
> steal inactive pages, which are pages which haven't been accessed for
> a pre-defined and tunable interval. Swapping is a desperation measure
> which will throw out the whole inactive process when the memory is
> at scarce.
>
> >
> > When the servers restarts , for a certain times , the swap usage is 0.
> > But i have the processes running for instance oracle , does this mean
> > that all the process executable is in memory?
>
> No. Page stealing (actually, it's called page replacement) will throw
> out inactive pages which were active only during the initialization phase
>
> >
> > Mladen , as you see i am little bit confused. Starting from the high
> > level aio question , i at last have the low level virtual memory
> > question.
>
> It's quite an extensive topic and there are several good Unix internals
> books on the market which will explain in detail how things work.
Received on Fri Jul 30 2004 - 00:49:15 CDT