RE: Jumpstart boot server with several interfaces: problem if client is not seen through primary interface

From: Christiaan Meihsl (Christiaan.Meihsl_at_reuters.com)
Date: 03/30/05

  • Next message: John Christian: "convert installed package back to package file?" 
    Date: Wed, 30 Mar 2005 16:27:52 +0200
To: sunmanagers@sunmanagers.org

    Hi all,

    Problem solved, thanks to inputs from Coy, Iain, Angelos and Carl. Thank
    you all for the great help !

    Summary of the solution:

    Jumpstart client "src1" was declared on 10.4.19 in /etc/hosts with
            10.4.19.51 src1
    And sysidcfg with
            {hostname=src1
            default_route=10.4.19.2
            ip_address=10.4.19.51

    BUT I was doing my boot net connected to the jumpstart interface with
    IP=192.168.1.21

    What is required is that the client have an IP declared in the same IP
    subnet as that
    where it contacts the jumpstart server.

    On jumpstart I changed /etc/hosts:
            192.168.1.51 src1
    And sysidcfg with
            {hostname=src1
            default_route=192.168.1.21
            ip_address=192.168.1.51

    And off it went !

    The error may seem obvious, but in my case the client is a blade, which
    will have, through the blade switch, 2 VLANS on 10.4.19 and 10.4.20.
    But if I understood correctly, the blade switch VLANs are not all
    accessible at the install stage, so I'm booting it through the blade
    switch VLAN 1 (on an untagged interface).
    It is on this VLAN that I plugged qfe1 of my jumpstart server
    192.168.1.21 ...
    So it was not so obvious, at least for me (not used to working with
    blades yet ...)

    Cheers,

    Christiaan
    ________________________________

    From: Christiaan Meihsl
    Sent: 30 March 2005 11:15
    To: 'sunmanagers@sunmanagers.org'
    Subject: Jumpstart boot server with several interfaces: problem if
    client is not seen through primary interface

    Hi gurus,

    I have a jumpstart server connected to several subnets, using the
    motherboard interface hme0, and a quad.
    hme0 is the primary interface, qfe0 and qfe1 are the other active (and
    tested) interfaces.

    Jumpstart works fine if the client connects to it via hme0 (primary
    interface).
    Jumpstart does not work if the same client connects to it through qfe1.

    If I snoop qfe1 I see:
    OLD-BROADCAST -> (broadcast) RARP C Who is 0:3:ba:2a:2c:28 ?
    This occurs each time I see "Timeout waiting for ARP/RARP packet" on the
    client.

    This looks like a previous problem (cf
    http://www.sunmanagers.org/pipermail/sunmanagers/2002-May/013553.html),
    except that the solution does not work ! I tried rebooting the jumpstart
    server, and manually restarting in.rarpd -a, to no avail.

    Jumpstart server is Ultra 5 with Solaris 8.
    Client is a blade SF B100 inside an SF B1600.

    Any suggestions ?

    Thanks in advance,

    CHRISTIAAN MEIHSL
    E-PSG - Senior Solutions Specialist
    Reuters SA

    REUTERS.KNOW.NOW.
    www.reuters.com <http://www.reuters.com/>
    www.reuters.com/customers <http://www.reuters.com/customers>

    -----------------------------------------------------------------
            Visit our Internet site at http://www.reuters.com

    To find out more about Reuters Products and Services visit
    http://www.reuters.com/productinfo

    Any views expressed in this message are those of the individual
    sender, except where the sender specifically states them to be
    the views of Reuters Ltd.
    _______________________________________________
    sunmanagers mailing list
    sunmanagers@sunmanagers.org
    http://www.sunmanagers.org/mailman/listinfo/sunmanagers

  • Next message: John Christian: "convert installed package back to package file?"

    Relevant Pages