Discuss October 2009

discuss@lists.centos.org

338 participants
325 discussions

weird mac address
by Janez Kosmrlj 26 Oct '09

26 Oct '09

Hi, I have an interesting problem. I have a centos server connected to our dmz zone. For security reasons our network department limited the number of mac addresses that can connect to the port on the switch. And since we are running vmware server on top on centos and a windows virtual machine inside of it, the number of mac addresses that can connect to this port is two or the switch blocks all traffic on this port. And now comes the problem. The port blocks almost immediately. After some investigation we found out, that we have actually 3 addresses communicating from this port. One that belongs to eth-0 and is ibm specific (the server is an IBM x3550 and every mac starts with the vendor code), one that belongs to the virtual machine and is Vmware specific and then we have another mac address that is IBM specific. I can't find where this traffic comes from. I traced the traffic with Wireshark and found, that this mac address creates a DHCP discover every 64 seconds. I first checked if this mac adress could be from the second network card, that is not connected with a cable nor is the device enabled, but it doesn't and the mystery mac is completely different from the network cards macs on the last 6 characters (the two on-board cards macs are different only on the last character). Then I tried searching the logs and /etc for any occurrence of this mac address, but I can't find anything. Then we tried to isolate the server, so I connected it directly to my laptop and run the trace again to see if this packet comes from outside and is then redirected back out on Vmware's virtual switch (since it's a broadcast packet, this could be somehow possible). But everything indicates that this packets are generated on this server. In the meantime we increased the number of mac addresses to 3, but after a while we got a packet from a fourth IBM mac address, that again doesn't belong anywhere. It was similar to the mystery mac (it was only different on the last two characters) and the switch blocked the port in the middle of the night. but this was a one-time occurrence for now. Does anyone have an idea what is happening or where i should look for the mystery mac address. thanx Janez

1 0

XFS might be available for i686 from the CentOS Plus repository.
by Rainer Traut 26 Oct '09

26 Oct '09

Hi, the topic is an excerpt of the 5.4 release notes and raises a question. I have a i386 server here, that has been running for nearly two years with xfs module from extras repo. When lookin at the module it was build against 2.6.18-92.1.13.el5. Question is, is it still safe to use this module with 5.4? Or should I switch back to a CentosPlus kmod and an older kernel? Rainer

2 2

OT: pager pay
by Alan McKay 26 Oct '09

26 Oct '09

Hey folks, I'm trying to compare my small company's policy against other companies out there. We pay 10% of your per-hour salary equivalent per every hour you carry it, whether it goes off or not. So for example someone making $55K/year works out to about $27 / hour (I think), so they'd get $2.70 for every hour they carry the pager outside of regular work hours. We are a company of about 25 people. My old employer Nortel still does it this way : - $27.50 for every 8 hour shift you carry it, outside of regular work hours. - any page is billed at time-and-a-half, minimum 1 hour billed - any page requiring you to go into work, is billed at time-and-a-half with a 3 hour min I have a buddy at IBM and while it varies by group, his group does this : - $30/day for carrying it, whether it goes off or not - every page gets billed at time-and-a-half, minimum 30 minutes Anyone else want to share theirs? thanks, -Alan -- “Don't eat anything you've ever seen advertised on TV” - Michael Pollan, author of "In Defense of Food"

7 7

CentOS DS Transition Issue
by Camron W. Fox 25 Oct '09

25 Oct '09

Alle, We are looking to transition from OpenLDAP to CentOS DS. We currently using OpenLDAP V2.3.43-3 with both RHEL 5.4 and Solaris10 clients (auth/netgroup/automount). We restrict user login access to the RHEL and Solaris machines using netgroups. After configuring the CentOS DS, we noticed that while the RHEL restrictions using the /etc/pam.d/system-auth-ac and /etc/security/access.conf configurations worked correctly, the Solaris clients using +@netgroup in /etc/passwd, passwd_compat in /etc/nsswitch.conf and the appropriate /etc/pam.conf configs do not respect the restrictions. The client configurations have not changed, only the LDAP server. I know this is a long shot, but has anyone else come across this issue? Best Regards, Camron -- Camron W. Fox Hilo Office High Performance Computing Group Fujitsu Management Services of America, Inc. E-mail: cwfox(a)us.fujitsu.com

1 0

Switch to measure traffic at IP level?
by Neil Aggarwal 25 Oct '09

25 Oct '09

Hello all: Is there a network switch that will give me traffic stats at an IP address level? Right now, I only get statstics at a port level, but that does not help since each of my servers run several virtual machines and I need to measure traffic per virtual machine. Thanks, Neil -- Neil Aggarwal, (281)846-8957, www.JAMMConsulting.com Will your e-commerce site go offline if you have a DB server failure, fiber cut, flood, fire, or other disaster? If so, ask about our geographically redundant database system.

12 32

permission script
by Eugeneapolinary Ju 25 Oct '09

25 Oct '09

I just wrote a little script, that will echo a message to the user, if it doesn't have enough permission: $ if ! [ "$(env LC_MESSAGES=C ifconfig wlan0 down | grep -i denied)" == "" ]; then echo "no permission"; fi SIOCSIFFLAGS: Permission denied $ sudo su [sudo] password for USER: # if ! [ "$(env LC_MESSAGES=C ifconfig wlan0 down | grep -i denied)" == "" ]; then echo "no permission"; fi # but it won't works why?

2 1

Which /usr/bin/qemu-img to use?
by Camron W. Fox 25 Oct '09

25 Oct '09

Alle, We're trying to install kvm on a newly update CentOS5.4 box. Which package should /usr/bin/qemu-img come from, kvm-qemu-img or qemu? When we try to install kvm\* we get the following error: Transaction Check Error: file /usr/bin/qemu-img from install of kvm-qemu-img-83-105.el5.x86_64 conflicts with file from package qemu-0.9.0-4.x86_64 file /usr/share/man/man1/qemu-img.1.gz from install of kvm-qemu-img-83-105.el5.x86_64 conflicts with file from package qemu-0.9.0-4.x86_64 Best Regards, Camron -- Camron W. Fox Hilo Office High Performance Computing Group Fujitsu Management Services of America, Inc. E-mail: cwfox(a)us.fujitsu.com

2 3

Testing remote machine status in a script ?
by Niki Kovacs 25 Oct '09

25 Oct '09

Hi, I'd like to test the status of a remote machine in a script. Something like: IF remote machine is switched on / connected to the network... ... THEN launch backup. How would you go about that? Cheers, Niki

4 7

Packages in 5.4 older than 5.3 updates, and no 5.4 updates available?
by Stephen Harris 25 Oct '09

25 Oct '09

I've been checking the updates tree for 5.4 and notice that they're empty. I checked 4 or 5 different servers and got the same results. This is, in my experience, unusual; there's typically patches for the new release from day one. But it could be that the 5.4 DVD is up to date. So I checked one of my machines and what updates it would need against 5.3 vs those that it would need to update to 5.4 (I've not been keeping my patches up to date on this machine; I know... stupid me). My machine is 33 patches short when configuring yum to a 5.3 tree (base and updates; it'd take 202 patches to take it up to 5.4). Of those 33 patches, 19 are the same from the 5.4 DVD. Of the remaining 14 discrepencies 10 had newer versions on the 5.4 disk. So far, so good. But 4 packages concern me where it looks like the 5.4 DVD has older versions than the 5.3 updates: 5.3 update:dnsmasq.i386 2.45-1.1.el5_3 5.4 DVD:dnsmasq.i386 2.45-1.el5_2.1 5.3 update:fetchmail.i386 6.3.6-1.1.el5_3.1 5.4 DVD:fetchmail.i386 6.3.6-1.1.el5 5.3 update:gnutls.i386 1.4.1-3.el5_3.5 5.4 DVD:gnutls.i386 1.4.1-3.el5_2.1 5.3 update:nspr.i386 4.7.5-1.el5_4 5.4 DVD:nspr.i386 4.7.4-1.el5_3.1 (These are the 4 my machine flagged; there could be other packages of similar concern that haven't hit me yet) Is this merely a case that the 5.4 updates tree hasn't yet been pushed out and these changes should show up in the near future (when?), have I found a problem, or am I just plain and simple wrong? :-) -- rgds Stephen

2 1

Running SSH on a different port (with SELinux)
by Jorge Fábregas 25 Oct '09

25 Oct '09

Hello everyone, Now after the recent discussion on running SSH on a different port, I decided to start a new thread but with SELinux involved. Assuming that you have SELinux enabled, and that you changed the default port for SSHD, let say for 1234, when I restart SSHD I don't get any AVC denials. This is the output of: semanage -l port | grep ssh ssh_port_t tcp 22 I thought (based on previous SELinux readings) that in order to allow SSHD on a non-default port you needed to: semanage port -a -t ssh_port_t -p tcp 1234 That was the theory I read :) Now in practice it seems it is not implemented yet, or at least by the time RHEL5 came out. Does anyone knows? All the best, Jorge

2 2

← Newer
1
...
7
8
9
10
11
12
13
...
33
Older →

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 October 2009