OSEC

Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com
 
From: Joshua E Warchol (jwarcholdsl.net)
Date: Tue Jun 05 2001 - 09:02:09 CDT

  • Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

    I've got to stop posting to the list before I check the docs! :-) That's
    twice recently I could have answered my own questions. Thanks!

    On Mon, Jun 04, 2001 at 02:45:59PM -0400, Wietse Venema wrote:

    > See the HISTORY file.
    >
    > Wietse
    >
    > Joshua E Warchol:
    > > Was that fixed in the lasted snapshot, I saw a patch hit the list
    > > and you said it was incorperated. I've already upgraded to
    > > Snapshot-20010525.
    > >
    > > On Sat, Jun 02, 2001 at 11:21:12AM -0400, Wietse Venema wrote:
    > >
    > > > Someone found a memory leak in the LDAP client a few weeks ago.
    > > > This explains why your memory footprint grows without bound.
    > > >
    > > > Wietse
    > > >
    > > > Joshua E Warchol:
    > > > > On Fri, Jun 01, 2001 at 02:21:28PM +0200, Rask Ingemann Lambertsen wrote:
    > > > > > > May 30 17:00:38 swangold postfix/virtual[13839]: A2749191A90:
    > > > > > > to=<seliaseliasproperties.com>, relay=virtual, delay=2323,
    > > > > > ^^^^^^^^^^
    > > > > > > status=deferred (maildir delivery failed: error writing message:
    > > > > > > Input/output error)
    > > > > > >
    > > > > > > I only get this on one server (it's part of a cluster, 10+ servers).
    > > > > >
    > > > > > Is it always the same message that fails?
    > > > >
    > > > >
    > > > > Yes, always exactly the same. I've been tracking the error for a few days
    > > > > now and am convinced it's NFS in the kernel being killed by layer-1 problems
    > > > > at the site. I'm going to upgrade to 2.4.5 and also have the cables replaced.
    > > > >
    > > > > On a side note I'm also fighting a memory problem as well. I've got postfix
    > > > > doing LDAP queries against a local slapd process, and then delivering the
    > > > > mail to a NFS mounted mail-store. Memory usage on the systems is almost
    > > > > reaching their limit (these have 512mb). Most of that appears to be 2.4.2
    > > > > agressively caching the NFS requests, on the order of 250-300MB. This
    > > > > shouldn't be a problem, as the kernel should give memory to processes that
    > > > > need it, but that doesn't seem to be the case. After a time the system will
    > > > > stop responding to ssh, then to console, then to ping, and finally it
    > > > > will stop responding entirely.
    > > > >
    > > > > I've got an _identical_ system running all the same software, but booted on
    > > > > 2.2.17, and memory usage is constant, and never reaches even 300MB. I'll be
    > > > > upgrading courier-imap and the kernel in the next few days to try to stop
    > > > > the problem. If you have a suggestion, feel free to drop me a line.
    > > > >
    > > > > Even with the few problems, its a tribute to postfix that a large email
    > > > > cluster can be built so easily! Go postfix!
    > > > >
    > > > > --
    > > > > Joshua Warchol
    > > > >
    > > > > -
    > > > > To unsubscribe, send mail to majordomopostfix.org with content
    > > > > (not subject): unsubscribe postfix-users
    > > > >
    > > > >
    > > >
    > >
    > > --
    > > Joshua Warchol
    > > UNIX Systems Administrator
    > > DSL.net
    > >
    > > -
    > > To unsubscribe, send mail to majordomopostfix.org with content
    > > (not subject): unsubscribe postfix-users
    > >
    > >
    >
    >
    > -
    > To unsubscribe, send mail to majordomopostfix.org with content
    > (not subject): unsubscribe postfix-users

    -- 
    Joshua Warchol
    UNIX Systems Administrator
    DSL.net
    

    - To unsubscribe, send mail to majordomopostfix.org with content (not subject): unsubscribe postfix-users