Udpcast February 2012

udpcast@udpcast.linux.lu

3 participants
5 discussions

exit code not passed if rx pipe fails
by George Coulouris 25 Apr '12

25 Apr '12

Hello, If udp-receiver is writing to a pipe and the pipe fails, udp-receiver will incorrectly exit with a returncode of zero. This is easily tested with a trivial pipe program that copies stdin to stdout, but exits with a nonzero status upon encountering eof. Adding the following to startReceiver() in udpr-negotiate.c is sufficient to fix this behavior: <---- cut here int ret; ... if(pipePid) { ret=udpc_waitForProcess(pipePid, "Pipe"); } ... return ret; <---- cut here Regards, George Coulouris National Center for Biotechnology Information (contractor)

2 1

Giving a talk on udpcast
by Knight Samar 10 Feb '12

10 Feb '12

Hi, I am giving a talk on UDPcast this year in GNUnify (http://gnunify.in) Any advice ? Regards, Samar

1 0

Gigabit switch question?
by Michael D. Setzer II 03 Feb '12

03 Feb '12

I got a Dell powerconnect 2724 gig switch, and have been testing it, but have not gotten the results I had hoped for. Have been trying a number of things, and only about to get speeds similar to using a 100M switch?? When transfering with 1 sender and 1 receiver I can get speeds of between 300M and 400M with no retransmits. Set max bitrate to 400M to avoid timeout messages. Have done that with two different clients with same results. When sending to the same two clients at the same time (or sending to 19 clients in another test) I only can get a speed of about 70M to 78M at best. With no options, it drops to 1M to 2M. Setting the max bitrate seems to reduce the rexmits, but the setting of 80M. I've tried it with --fec 8x8/128 and without, and doesn't seem to make much difference. It might be the something in the switch, but I have gone in, and disable the storm protection settings that I had though might be an issue early in the process? The switch does have some limitations. Had to use an old 98 machine with old firefox to access its web interface, since it doesn't seem to work with newer browsers. Might just be this switch doesn't do well with multi-cast. +----------------------------------------------------------+ 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 G4L Disk Imaging Project maintainer http://sourceforge.net/projects/g4l/ +----------------------------------------------------------+ 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 11810129.071646 | EINSTEIN 7213450.559851 ROSETTA 4201950.390797 | ABC 10802915.987447

1 0

Drivers for NIC and HDD
by Rahul Rao 02 Feb '12

02 Feb '12

I was just looking out for drivers for NIC and Hard Disk. Their model numbers are as follows, NIC: 1) Realtek RTL8168C(P)/8111C(P) Family PCI-E Gigabit Ethernet NIC(NDIS 6:20) Harddisk: 1) WDC WD2500 AAJS-60M0a0 ATA DEVICE 2) Hitachi HDP725016GLA380 ATA DEVICE Can you suggest me any other driver which can be compatible with this model of the hardware. -- Regards, Rahul Rao Msc-CA, SICSR

2 1

using udpcast with 1G ethernet.
by Michael D. Setzer II 01 Feb '12

01 Feb '12

I just got a 24 port 1GB switch and my first attempt definitely didn't give the results I was hoping for. Perhaps there are some tuning options that might improve the options. First, I created an ntfsclone lzop compressed image of the Windows partition that is uploaded to one of the machines connected with the 1GB switch. That took about 11 minutes. I then made sure that all the machines had the nics running at 1G. Originally, two machines were only connecting at 100M, but replacing the connector on 1 and just recrimping the other got them all to connect at 1G. The cables are all Cat5e, but they are a few years old. I have a script that using udpcast to transfer this image file to all the other 19 machines in the lab connected to the machine. I was hoping to get a better speed when switching to the 100M to the 1G, but wasn't looking for a 10 times increase, since I knew that the hard disks only will do about 55MB/ second accordng to hdparm. But in my first test it took 31 minutes to transfer the file using udpcast?? Image file is about 12GB. Here is the end of the listing. bytes= 12 221 254K re-xmits=32506028 (378.1%) slice=0112 - 15 Timeout notAnswered=[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18] notReady=[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18] nrAns=0 nrRead=0 nrPart=19 avg=252 bytes= 12 229 698K re-xmits=32529469 (378.2%) slice=0112 - 3 Transfer complete. Didn't notice any machine giving errors, since it seemed to always list all 19 machines. The re-xmits seems to be an issue, but perhaps some other setting with the Linux (Fedora 16) or udpcast options. The script is run from the machine machine, and it links to the other 19 machines and then transfers. Partition /dev/sda6 is is a ext3 partition that is separate on all machines to store images, that can then be used to reimage individual machines with no network traffic. if [ -e /sda6/d7-xp.lzo ] ; then echo "SDA6 already mounted" else sudo mount /dev/sda6 /sda6 fi cmd="hostname && mount /dev/sda6 /sda6 && cd /sda6 && ./udp-receiver --no-progress --file d7-xp.lzo && cd / && umount /sda6 " #cmd="udp-receiver --file d7-xp.lzo " plink -t 192.168.7.201 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.202 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.203 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.204 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.205 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.206 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.207 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.208 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.209 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.210 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.211 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.212 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.213 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.214 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.215 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.216 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.217 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.218 -l root -pw xxxxxx "$cmd" & plink -t 192.168.7.220 -l root -pw xxxxxx "$cmd" & sleep 20 cd /sda6 time udp-sender --full-duplex --file d7-xp.lzo --min-receivers 19 --min-wait 9 I'm hoping there are some more experience with using udpcast in 1G environments. Currently running another test with sending to just two machines, but seems be showing similar results. +----------------------------------------------------------+ 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 G4L Disk Imaging Project maintainer http://sourceforge.net/projects/g4l/ +----------------------------------------------------------+ 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 11801304.490587 | EINSTEIN 7206065.819851 ROSETTA 4194793.509365 | ABC 10747635.646984

2 2

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

Udpcast February 2012