This was working and now it has stopped.
Have a UDPcast server running on a HP DL360G5 with 4 NICS. One
connected to an internal infrastructure (eth2) and another connected
to a dedicated switch with the client systems on the same switch (eth3).
When i run the UDPcast send or receive, using the following command:
udp-sender -f /home/base.udpc --interface eth3 --point-to-point --full-
duplex --max-bitrate 600m --nokbd --portbase=9922 --min-receivers 1 --
rexmit-hello-interval 5000 --max-wait 300
I get the following error message:
Udp-sender 2007-03-23
No suitable network interface found
The following interfaces are available:
lo 127.0.0.1
eth2 10.46.0.251
eth2:1 10.46.0.1
Interestingly, an ifconfig show eth3 as a valid network adapter, but
the udp-sender does not see the device. This was working last week,
and nothing has changed on the server
I have tried everything, restarting DHCP, removing the interface and
adding again. Rebooted the server, restarted the network stack
manually. nothing seems to help.
Using SUSE linux 10 as the base server OS.
Hello,
I've got problem with multicasting to many computers at once. I've
got server with newest udp-sender (20071228) which is sending files from
ext3 partition (with command ./udp-sender --pipe 'tar -B -S -cpf - -C
/usr/share/systemimager/boot/i386/pjwstklenny .' --portbase 9006
--min-clients 2 --min-wait 15 --interface eth1 --max-bitrate 70M
--mcast-all-addr 234.1.1.162 --ttl 10
and I'm trying to receive it with command:
./udp-receiver --portbase 9006 --nosync --ttl 10 --mcast-rdv-address
234.1.1.162 --pipe "tar -f - -x -vv -C /disk"
on client computer.
Problem is that, when I'm doing it one-by-one (I mean server-1 client) I
get full speed - ~70MB/s, but when I'm trying to to this with >= 2
computers transfer drops to >0,5Mb/s. Why? I tried diffrent and the same
subnet, I'm using 3com & cisco 3-layer switches (I also tried putting
server and clients on the same switch - the same effect).
Please help, cause I've got >200 computers to clone ;)
Best regards
Paul Ufnalewski // BSS
Polish Japanese Institute of Information Technology
Both server (debian etch) and client (knoppix 5.1 with downloaded and
compiled new version of udpcast) don't have firewalls. Maybe it is
something with kernel version (on both server and client)? I will try
udpcast boot cd...
Michael D. Setzer II pisze:
> Just a follow up to the other reply you have already gotten. If you have a firewall running on
> the client and servers, you need to open both the 9006 port and the 9007 ports. Not sure if
> this would effect you. I was sending a file to many machine, and it also worked with one
> client, but more, it didn't work. I think the transfer is done on the main port, but the clients talk
> back to the sender machine via the next port. I also had the same switch problem, but using
> udpcast boot cds don't have the firewall, but the Fedora systems do.
>
>
>
> On 23 Sep 2008 at 22:05, Paweł Ufnalewski // BSS wrote:
>
> Date sent: Tue, 23 Sep 2008 22:05:25 +0200
> From: Pawe³ Ufnalewski // BSS <archon(a)pjwstk.edu.pl>
> To: udpcast(a)udpcast.linux.lu
> Subject: [Udpcast] multi computer cast problem
>
>
>> Hello,
>>
>> I've got problem with multicasting to many computers at once. I've
>> got server with newest udp-sender (20071228) which is sending files from
>> ext3 partition (with command ./udp-sender --pipe 'tar -B -S -cpf - -C
>> /usr/share/systemimager/boot/i386/pjwstklenny .' --portbase 9006
>> --min-clients 2 --min-wait 15 --interface eth1 --max-bitrate 70M
>> --mcast-all-addr 234.1.1.162 --ttl 10
>>
>> and I'm trying to receive it with command:
>>
>> ./udp-receiver --portbase 9006 --nosync --ttl 10 --mcast-rdv-address
>> 234.1.1.162 --pipe "tar -f - -x -vv -C /disk"
>>
>> on client computer.
>>
>> Problem is that, when I'm doing it one-by-one (I mean server-1 client) I
>> get full speed - ~70MB/s, but when I'm trying to to this with >= 2
>> computers transfer drops to >0,5Mb/s. Why? I tried diffrent and the same
>> subnet, I'm using 3com & cisco 3-layer switches (I also tried putting
>> server and clients on the same switch - the same effect).
>>
>> Please help, cause I've got >200 computers to clone ;)
>>
>> Best regards
>>
>> Paul Ufnalewski // BSS
>> Polish Japanese Institute of Information Technology
>>
>>
>> _______________________________________________
>> Udpcast mailing list
>> Udpcast(a)udpcast.linux.lu
>> https://lll.lgl.lu/mailman/listinfo/udpcast
>>
>
>
> +----------------------------------------------------------+
> Michael D. Setzer II - Computer Science Instructor
> Guam Community College Computer Center
> mailto:mikes@kuentos.guam.net
> mailto:msetzerii@gmail.com
> http://www.guam.net/home/mikes
> Guam - Where America's Day Begins
> +----------------------------------------------------------+
>
> http://setiathome.berkeley.edu (Original)
> Number of Seti Units Returned: 19,471
> Processing time: 32 years, 290 days, 12 hours, 58 minutes
> (Total Hours: 287,489)
>
> BOINC@HOME CREDITS
> SETI 6,443,581.5070 | EINSTEIN 1,996,973.6209 | ROSETTA 618,619.3627
>
>
Is it possible to do dynamically throttled UPDcast transmission? We need
to throttle the data flow to a Logic Innovations IP encapsulator (with
LI's documentation attached with their permission).
The udpcast mutlticast address is considered a single route to the IPE.
The IPE sends out flow-control messages (described in document) that
shows how many bytes are available for traffic for each route every 100
milliseconds in a multicastand comes from a user specified multicast of
its own. Attached is the LI document, again with the manufactures
permission. Also, you can contact Theo Aukerman, taukerman(a)logici.com
for any additional info.
I'm asking these questions because IF this is possible, then this would
be an almost complete framework for an open source Data-casting network,
and can be used for over the air data-broadcasting from any DTV
transmitter or any other system with null packets. I currently work as
an engineer for a TV station that does data transmission, but would like
to move to an open source model.
Hi everyone,
Please confirm my guess (or explain how to implement it).
I need a "sender-timeout", a period of time while sender waits for receivers
to connect, and if none are connected - sender exits without transmission.
The only way to implement this is shell/perl/... script, which starts
udp-sender with --min-receivers option and monitors its stderr. If no
"Starting transfer" is found during specified period of time - it kills
udp-sender process.
Regards,
Dmitry
Hi,
======================
I already sent this message, but I get troubles using the other email...
======================
I don't succeed in giving two options through apend boot cmdline.
Here is my cmdline:
label u1auto
kernel vmlu26
append initrd=udprd root=01:00 persoparams=oui lang=FR kbmap=FR dhcp=yes
compr=lzop port=9002 umode=snd disk=/dev/hda1 auto_reboot=always
enableDiskmodule=yes diskmodule=AUTO netmodule=AUTO
udpcparam=--max-wait=300--min-receivers=2
A script is launched using predialog.
It writes the /udpcfg.txt file.
With the above cmdline, I get:
udpsender -p lzop --file /dev/hda1 -P 9002 -i
eth0 --max-wait=300--min-receiver=2
instead of
udpsender -p lzop --file /dev/hda1 -P 9002 -i
eth0 --max-wait=300 --min-receiver=2
And it doesn't work.
===========================
I've tried:
label u1auto
kernel vmlu26
append initrd=udprd root=01:00 persoparams=oui lang=FR kbmap=FR dhcp=yes
compr=lzop port=9002 umode=snd disk=/dev/hda1 auto_reboot=always
enableDiskmodule=yes diskmodule=AUTO netmodule=AUTO
udpcparam=--max-wait=300|--min-receivers=2
And my predialog script does:
udpcparam=`echo "$udpcparam" | sed -e "s/|/ /g"`
echo "udpcparam=$udpcparam" >> /udpcfg.txt
But it doesn't work.
/udpcfg.txt contains
...
umode=snd
udpcparam=--max-wait=300 --min-receivers=2
And I get
Udp-sender 2007-12-28
--min-receivers=2: not found
Using mcast address 232.168.52.210
Compressed UDP sender for /dev/hda1 at 192.168.52.210 on eth0
Broadcasting control to 192.168.52.255
and it fails.
How could I get both options working?
I did those tests using 2.6.23 kernel and udpcast-mkimage-20080107-1.i586,
udpcast-20070602-1.i586
I'll try again using
udpcast-20071228-1.i586
udpcast-mkimage-20080911-1.i586
kernel-udpcast-2.6.26.3-1.i586
Regards.
--
Stephane
I have some more information on the error. Instead of using a
pre-configured cast-o-matic initrd, I walked through the boot process
and at the step where the tg3 driver was loaded I received a kernel
message, part of it looks like this.
PCI: Setting latency timer of device ...
BUG: Unable to handle kernel NULL pointer dereference at 0000001c
IP: [<f8857057>]
*pde: 0002 [#1]
Modules linked in: tg3(+) ...
I don't know if it will but I hope this helps.
Scott Mikolyski wrote:
> Alain Knaff wrote:
>> Scott Mikolyski wrote:
>>> Hi,
>>>
>>> I'm working with a pxe server to get workstations to boot from the
>>> network. I'm using the cast-o-matic boot image.
>>>
>>> Everything works up until I get and error message "ifconfig:
>>> SIOCSIFADDR: No such device"
>>>
>>> Dropping to a shell, I type ifconfig and get nothing. I think I
>>> should be getting some nic information.
>>>
>>> So it looks like to me that the network driver is not getting loaded.
>>> I've selected different modules in cast-o-matic. But I feel that it
>>> has something to do with this version of the kernel.
>>>
>>> Last month I had this setup working perfectly. Since then, I've lost
>>> my pxe server and need to recreate it. Is it possible that this
>>> recent change broke something?
>>>
>>> Please help. Thank you.
>>
>> Could you tell me which NIC it is (name of module which worked in old
>> kernel), so that I can investigate? Or, lacking that, the output of lspci
>
> It's a Broadcom NetXtreme BCM5751.
>
>>
>> Thx,
>>
>> Alain
>>
>> _______________________________________________
>> Udpcast mailing list
>> Udpcast(a)udpcast.linux.lu
>> https://lll.lgl.lu/mailman/listinfo/udpcast
>
Hi,
I'm working with a pxe server to get workstations to boot from the
network. I'm using the cast-o-matic boot image.
Everything works up until I get and error message "ifconfig:
SIOCSIFADDR: No such device"
Dropping to a shell, I type ifconfig and get nothing. I think I should
be getting some nic information.
So it looks like to me that the network driver is not getting loaded.
I've selected different modules in cast-o-matic. But I feel that it has
something to do with this version of the kernel.
Last month I had this setup working perfectly. Since then, I've lost my
pxe server and need to recreate it. Is it possible that this recent
change broke something?
Please help. Thank you.
