Drop NFS speed on move from kernel 2.4 to 2.6 (was: Change in NFS client behavior)

From: Tomasz Kłoczko (kloczek_at_rudy.mif.pg.gda.pl)
Date: 09/02/05

  • Next message: Alexey Kuznetsov: "Re: Possible BUG in IPv4 TCP window handling, all recent 2.4.x/2.6.x kernels" 
    Date:	Fri, 2 Sep 2005 17:06:20 +0200 (CEST)
To: Rob Sims <lkml-z@robsims.com>

    On Wed, 31 Aug 2005, Rob Sims wrote:

    > We have noticed when changing from kernel 2.4.23 to 2.6.8 that
    > timestamps of files are not changed if opened for a write and nothing is
    > written. When using 2.4.23 timestamps are changed. When using a local
    > filesystem (reiserfs) with either kernel, timestamps are changed.
    > Symptoms vary with the client, not the server. See the script below.
    >
    > When run on a 2.4.23 machine in an NFS mounted directory, output is
    > "Good." When run on a 2.6.8 or 2.6.12-rc4 machine in an NFS directory,
    > output is "Error."
    >
    > Is this a bug? How do we revert to the 2.4/local fs behavior?

    I have another (IMO strange) observation on move NFS client from kernel
    2.4. to 2.6: drop NFS speed by *~1/3*.

    In attachemt is small giff generated by mrtg from data sucked from
    switch SNMP agent. Each day (at night) I perform dayly backups (via NFS)
    and on this plots it visable by dayly high network activity.
    Aprox four weeks ago I'm switch on host connected to monitored switch port
    from kernel 2.4 to 2.6. Result can be observer on attached picture by
    reduce maximum bandwitch data transferred on this port by ~1/3 (?!?).

    Next smaller drob down (~two weeks ago) which can be observed on this plot
    it is result switchung from UDP to TCP (performed specialy for produce
    some visable data for compare).

    ~30% drob down NFS client speed on this host isn't critical but probably
    this is kind of top iceberg with some other problems which now exist in
    kernel 2.6. I'm not expert in this area but my frient (Marcin Dalecki)
    says this drop down may have some roots in worse memory management in
    kernel 2.6. Is it can be true or not ?

    kloczek 

    -- 
-----------------------------------------------------------
*Ludzie nie mają problemów, tylko sobie sami je stwarzają*
-----------------------------------------------------------
Tomasz Kłoczko, sys adm @zie.pg.gda.pl|*e-mail: kloczek@rudy.mif.pg.gda.pl*

    -
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/

  • Next message: Alexey Kuznetsov: "Re: Possible BUG in IPv4 TCP window handling, all recent 2.4.x/2.6.x kernels"

    Relevant Pages