Discuss

discuss@lists.centos.org

34758 discussions

Fwd: RHEL5 network throughput/scalability
by Walid 13 Jun '08

13 Jun '08

Hi All, I have posted this messgage yesterday in the Beowulf mailing list, and did not get any responses, as i have tried different Centos kernels to see if the behaviour changes or not, and it did not change much, I am posting it here, I hope no one minds, and thanks in advance for any pointers or clues? ---------- Forwarded message ---------- From: Walid <walid.shaari(a)gmail.com> Date: 2008/6/12 Subject: RHEL5 network throughput/scalability To: beowulf(a)beowulf.org Hi All, I have an issue with a new cluster setup where the nodes are RHEL5.1(with the latest 5.2 kernel), when i try to write NFS data, the nodes scale linearly until they reach the 10th node, that is the bandwidth , and throughput seen from the NFS sever on the other side of the nodes shows a liner increment from around 100+Mbyte/sec up to 1Gbyte/sec, however when we add another extra node to the equation the bandwidth/throughput becomes erratic/inconsistent, and drops to around 500-700Mbyte/sec. however if i try the same setup with RHEL4U6 i do not get the same behaviour it sustains the bandwidth at 1Gbyte/sec. the setup is like this 48 nodes sharing 48 port access switch that is up linked using 10g link to a CISCO 6509 switch which is linked to a Clustered NFS File system that consist of eight heads where each head linked using a 10G link to the 6509. the above was a write test, so i thought may be the tcp congestion kicked in, or sliding windows problem, however when i do a read test it gets worse, the scalability now is reduced to 5 nodes that is one node is able to read around 100 MBps, two will read double, and so on until you add the fifth node where the bandwidth drops from around 500+MBps to around 300, and again from RHEL4 the behaviour is different. any pointers? TIA Walid

2 2

dependency on /usr/lib/nx
by drew einhorn 13 Jun '08

13 Jun '08

I do a: sudo yum update and I get: yada, yada, yada,... ---> Package freenx-server.i386 0:0.7.2-8.el5 set to be updated --> Running transaction check --> Processing Dependency: /usr/lib/nx for package: freenx-server Importing additional filelist information --> Finished Dependency Resolution Error: Missing Dependency: /usr/lib/nx is needed by package freenx-server Not only does this sound like an odd dependency /usr/lib/nx exists! Any ideas? -- Drew Einhorn

2 1

pthread_setschedprio function in glibc-2.3.4
by Dongyoon Lee 12 Jun '08

12 Jun '08

I re-installed glibc-2.3.4-2.39.src.rpm on CentOS 4.6, and tried to use pthread_setschedprio() API in glibc library, but it fails. Here is my system information. Operating System CentOS 4.6 (x86_64), Linux version 2.6.9-67.Elsmp gcc version 3.4.6 20060404 (Red Hat 3.4.6-9) glibc (pthread) src glibc-2.3.4-2.39.src.rpm I checked that setschedprio function is in the GLIB_2.3.4 but I got following errors [root@localhost TMP]# gcc d.c -o d -lpthread /tmp/ccEeiBLd.o(.text+0x179): In function `main': : undefined reference to `pthread_setschedprio' collect2: ld returned 1 exit status Can I get any advice to use that API?

1 0

raid1 disk format?
by Les Mikesell 12 Jun '08

12 Jun '08

If you have a disk with several partitions set up as members of a raid1 md devices, can you make a dd image of that disk to replace its matching drive with identical partitions or are there differences between the mirrored partitions? -- Les Mikesell lesmikesell(a)gmail.com

4 15

mdmonitor not triggering program on fail events
by Ow Mun Heng 12 Jun '08

12 Jun '08

Hi, Can anyone help me with this since I want to get it done "the correct way" I'm trying to make mdmonitor to execute a program when it detects a fail event automatically. Currently, from what I see, init is calling mdmonitor with these options mdadm --monitor --scan -f (note that the --program is not there) and this is in my /etc/mdadm.conf MAILADDR root PROGRAM /root/program_2_run.sh short of hacking the mdmonitor script to hardcode the program there, is there an alternate, more elegant way? Thanks.

4 6

OT: HP Autoloader Issue
by Joseph L. Casale 12 Jun '08

12 Jun '08

I spent a good chunk of time trying to get an HP LTO Autoloader to work with Amanda only to later find out it was faulty. Aside from HP who has a brutal system for repairing hardware, does anyone know where I might be able to send it for repair? Thanks! jlc

2 1

upgrade Fedora -> Centos
by Sergej Kandyla 12 Jun '08

12 Jun '08

Hi all, Is there some way to upgrade from Fedora Core 6 to Centos 5 (on remote serv) ? Thanks in advice! -- Best Wishes, PAIX-UANIC | SK3929-RIPE

4 3

Problems installing 5.1 on a Tyan Thunder HEsl with a SCSI controller
by Timothy Selivanow 12 Jun '08

12 Jun '08

Hi, I'm trying to install 5.1 using the onboard LSI Symbios 53C1010, and I'm running into some trouble. When the computer first boots, the SCSI BIOS sees the three HDDs, but when I go to install, the installer hangs for a while at inserting the sym53c8xx driver and if I go over to the screen on F4 it shows that it is trying to scan the SCSI bus and is resetting all of the IDs. Once that is done, it moves on the the actual installer, but does not see any drives. I'm kind of at a loss at the moment. I feel like there would be a kernel boot option that I could give the installer, but I don't know what that would be. I know the SCSI IDs of the three HDDs (and the controller), is there a way to say "just use these IDs and move along"? I'm currently downloading a Fedora 9 CD (I don't have a spare DVDROM) to see if a newer kernel will help, as going back down to CentOS 5.0 didn't (I had it at hand). Thanks for any help! --Tim

2 2

rebuilding kernel modules issues
by Kenneth Burgener 12 Jun '08

12 Jun '08

I have a Tyan S3870 (K855A) motherboard which I would like to perform a burn in test with. lm_sensors needs a kernel module compiled to be support this motherboard's sensors. The instructions, for compiling the module, by tyan (ftp://ftp.tyan.com/software/lms/3870.sensors.conf) match closely with the instructions I found here (http://www.howtoforge.com/kernel_compilation_centos) I simply installed the kernel source (yum install kernel-devel) and tried the first basic step, cleaning (which is probably not needed at this point), but it threw the first of many errors I have run into. # uname -a Linux fwtest 2.6.18-53.1.21.el5 #1 SMP Tue May 20 09:34:18 EDT 2008 i686 athlon i386 GNU/Linux # cd /usr/src/kernels/2.6.18-53.1.21.el5-i686 # make clean && make mrproper scripts/Makefile.clean:17: /usr/src/kernels/2.6.18-53.1.21.el5-i686/drivers/infiniband/hw/amso1100/Makefile: No such file or directory make[3]: *** No rule to make target `/usr/src/kernels/2.6.18-53.1.21.el5-i686/drivers/infiniband/hw/amso1100/Makefile'. Stop. make[2]: *** [drivers/infiniband/hw/amso1100] Error 2 make[1]: *** [drivers/infiniband] Error 2 make: *** [_clean_drivers] Error 2 I then proceed to the next step of "make menuconfig". I select the driver "i2c-piix4" and then save, exit, and attempt to compile the modules. [root@fwtest 2.6.18-53.1.21.el5-i686]# make modules CHK include/linux/version.h CHK include/linux/utsrelease.h make[1]: *** No rule to make target `arch/i386/kernel/msr.c', needed by `arch/i386/kernel/msr.o'. Stop. make: *** [arch/i386/kernel] Error 2 I tried deselecting the "i2c-piix4" driver, and tried again with the same results. I also tried grabbing the .config file from /boot/ (cp /boot/config-`uname -r` ./.config) with the exact same results. This is a clean, newly installed, fully updated system. Shouldn't I be able to do this without error? Am I missing another package to be able to compile the kernel modules as is? Thanks for any help, Kenneth

2 2

General CentOS 5.1 (or Gnome) instability?
by Alfred von Campe 12 Jun '08

12 Jun '08

I've been a big fan of CentOS for a while, and didn't have many issues with CentOS 4.X over the past few years. However, since moving to CentOS 5.1 a few weeks ago, I have received more problem reports from my users than in the last year and a half on CentOS 4.X. I've previously reported the problem with gnome-terminal crashing (and since there is a single gnome-terminal process by default all your terminal windows disappear which makes this really painful), and now I'm getting multiple reports of Gnome applets suddenly quitting. Sometimes this includes the entire screen "flashing" (probably a side effect of the "Show Desktop" applet exiting). I've also had reports of some third party tools like SlickEdit misbehaving and/or crashing on CentOS 5.1. Don't get me wrong; I'm not really complaining about CentOS. I really appreciate what the CentOS team does. I am just wondering if anyone else has seen these issues. If it was just one or two users, I would suspect the hardware or some configuration issues. But these issues started cropping up after I upgraded our existing systems to CentOS 5.1. The "upgrade" was a complete reinstall via a kickstart script (I reformatted all partitions/LVs except for one), and all systems are configured identically. BTW, I'm using CentOS for our desktops as well as our servers, and all these problems are really confined to the desktop systems. Almost everyone uses the default Gnome desktop. So, does anyone else have the perception that CentOS 5.X (particularly Gnome) is a little less stable than CentOS 4.X or is it just me? Alfred

6 9

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Discuss