This patch applies to 20070602 and has a few fixes needed for OS X 10.4.
* removes 3 harmless warnings in fifo.c, log.c, statistics.c, and
util.c
* drops the fatal error when getting the broadcast address fails
for an
interface, since this is normal for the loopback, which is
checked before
the ethernet. This causes udpcast to always die unless the --
interface
option is used.
* use the alternative IP_MULTICAST_IF method, same as for Cygwin.
The
comment in the code made it sound like Cygwin was out of the norm,
but it's Linux that's at ends with OpenBSD, Apple OS X, Solaris,
Cygwin, etc...
Thanks,
Devin Bayer
Hello. I have been using udpcast (version 2006-09-15) successfully
until last week when I decided to upgrade. I first tried 2007-06-02,
but that didn't work. So I tried 2007-03-26 and that failed in the
same way.
When I use either of these newer images I see this message on the
receiver:
...
sending go signal
lzop: <stdin>: block size too small -- recompile lzop
bytes= 848 848 (43.00 Mbps)
and then the transfer stops. The program doesn't end, but the sender
just continues to timeout.
Any help would be appreciated. It must have happened between Sept,
2006 and March, 2007. I am using the prebuilt images from the CD-
ROM, booted via PXE, as my receiver. My sender is Mac OS 10.4
PowerPC and I tried the Sept '06 and most recent versions of the
program. I tried lzop 1.0 and lzop 1.02rc2, both of which work with
the older receiver.
Thanks,
Devin Bayer
Hi everybody,
I have a problem with the images castomatic for PXE.
When I download the PXE image already prepared on the site, there is no
problem but if I want to create my own preconfigured image with
castomatic, the image will load but an error occures and we can see the
following error message :
File "/lib/modules/2.6.21.3udpcast/modules.pcimap" not found (No
such file or directory)
I tried 3 images :
- Preconfigured for a Dell optiplex 740 (NIC => tg3, Sata =>
sata_nv)
- Preconfigured for a Dell optiplex GX620 (NIC => tg3, Sata =>
ata_piix)
- Preconfigured (NIC => Autodetected, Sata =>Autodetected)
Please note that my DHCP server is on a Windows Server and that my TFTP
server is on another Windows Server. I don't think the problem come from
my infrastructure. The problem should come during the creation of the
image by castomatic.
Is someone already encounter this issue ?
Thanks a lot for any response,
--
Cyrille MAECHLER
IT Manager European School Luxembourg
23, Boulevard Konrad Adenauer
L-1115 Luxembourg
Phone : +352 432082-408
Fax : +352 432082-344
cmaechler(a)euroschool.lu
------------------------------------------------------------------------
-
###########################################
European School Luxembourg 1:
This message has been scanned by F-Secure Anti-Virus for Microsoft Exchange.
For more information, connect to http://www.f-secure.com/
Hi,
I've got in my school 40 pcs - amd 4200+ 1GB Ram 160 GB Samsung
IDE-harddisk and Intel pro 1000GT - Gbit-switch
If I clone a receiver and the sender directly with a crossovercable
I've got a transferrate of nearly 500 Mbit/s. If I clone at once 4 pcs
with the Gbit-switch I've got a transferrate of 350Mbit/s. If I clone
more then 5 pcs at once the transferrate falls down to 70 - 90 Mbit/s.
Well that's okay and udpcast is fantastic but if I could clone 20 pcs
also with 500 Mbit/s it would great.
Does anyone knows the reason for the reduced transferrate - whichever
could be the bottleneck?
Kind regards
Dieter
Hi;
Hopefully you can shed some light on the situation I am encountering. I
have a class set of 30 Dell Optiplex 745 workstations - and the drivers
on the Web Configurator (primarily the tg3 driver) did not work with the
Broadcom NetXtreme Gigabit network adapter in these workstations.
So I built a new set of disks using a Feisty Fawn live cd and its Kernel
to get the appropriate driver support. I am now able to get udpcast
working to send a sysprepped windows xp pro image, ntfs - 80gb sata
drive(ata-piix).
I am now having the issue where machines are randomly failing with the
newly casted image. I've been using lzop compression and have received
a few errors on the sender - but the my main concern is that random
machines get a disk read error (obviously wont boot) and some get NTLDR
Missing error messages. This is very odd as it isn't always the same
machines that get this error after a cast. I can restore a ghost image
to those machines individually - they then initialize just fine.
I've also tried to create disks using the Fedora Core 7 live cd with the
same results. I am attaching a the very rough documentation I made as a
reminder to myself on how I create the udpcast disks.
Udpcast is great and works flawlessly on my older Dell labs here in the
college!
The issue has stumped me for about a month as I've tried different
techniques on and off. Your insight is greatly appreciated!!
Alex Richards
IT Specialist, Information Technology Services
saultcollege.ca <http://www.saultcollege.ca/>
E-mail: alex.richards(a)saultcollege.ca
<mailto:alex.richards@saultcollege.ca>
make -j2 was failing because it started a make each for udp-receiver.1 and
udp-sender.1.
make DESTDIR=/tmp install was failing because the DESTDIR glue wasn't
included. This is a standard part of autoconf and really helps when staging
an install.
Finally updated the gentoo packages associated with udpcast.
http://packages.gentoo.org/packages/?category=net-misc;name=udpcast
Keep up the great work,
Cheers,
--
Daniel Black <dragonheart(a)gentoo.org>
Gentoo Foundation
by Stephane Boireau (Animateur TICE Bernay/Pont-Audemer)
Hi,
It seems, a trouble happens using busybox v1.5.0 (2007-03-28 14:34:43).
I build an iso using:
# udpcast-20070323-1.i586.rpm
# udpcast-mkimage-20070328-1.i586.rpm
# kernel-udpcast-2.6.20.4-1.i586.rpm
And I get strange things:
# date
Fri May 4 07:50:34 UTC 2007
#
That's OK.
But:
# date > /tmp/date.txt
# cat /tmp/date.txt
#
# A=`echo world`
# echo $A
world
#
That's OK.
But:
# A=`date`
# echo $A
#
No completion working.
Is it possible to correct this?
Thanks!
--
Stephane Boireau
Animateur TICE Bernay/Pont-Audemer