Discuss September 2004

discuss@lists.centos.org

60 participants
113 discussions

Re: RedHat/CentOS family tree
by Rick Graves 15 Sep '04

15 Sep '04

John, Your RedHat/CentOS family tree looks great! Even people who are not trying to migrate might want to take a look. For those who have not seen it yet, it is at the bottom of John's migration page, http://uranus.it.swin.edu.au/~jn/linux/centos-2/migration.htm For a general understanding how the different versions relate to each other, that picture is worth 1,000 words, easy. Thanks, Rick --- centos-request(a)caosity.org wrote: > Message: 2 > Date: Wed, 15 Sep 2004 10:35:21 +1000 > From: John Newbigin <jn(a)it.swin.edu.au> > To: seth vidal <skvidal(a)phy.duke.edu> > Cc: centos(a)caosity.org > Subject: Re: [Centos] Using yum to do downgrades > > Thanks for your reply. I guess for now I will stick > with my bash > script. I'll put it up on my web site > http://uranus.it.swin.edu.au/~jn/linux/centos-2/migration.htm > > John. >

1 0

RE: Getting kernel panic when installer starts formatting partitions. Was: [Centos] How to get 3Ware 9500S controller card working in CentOS 3.1?
by David Corredor 15 Sep '04

15 Sep '04

In regard of the e-mail sent on Jul/16/04 about the 3Ware 9500S. Has anyone experienced/solved this issue ?? I'm having a similar problem, I already have the system installed on a separate disk, now I want to make the filesystem on the RAID partition (using 3Ware 9500S-8). First I had the problem that the card will keep resetting after continous timeouts whenever I tried loading the module (compiled from source, in up-mode, driver v.9.2 from 3Ware's website). The solution to this was to disable APIC on the motherboard (or boot linux without APIC support). So, now the module loads ok (I've tried SMP and ENT versions, I'm using kernel 2.4.21-15.0.4-ELsmp). But the timeout and reset problem start happening again when I try to create the filesystem on the RAID partition. I can cancel the mkfs operation, but then I get a kernel panic if I try to unload the module (rmmod 3w-9xxx), just like the problem that the previous poster had. Any ideas or input would be appreciated. David

2 1

Using yum to do downgrades
by John Newbigin 15 Sep '04

15 Sep '04

This is a yum question but relates to some work I am doing with CentOS and other recent discussions on this list so I though I would post here first. Basically I am attempting a Red Hat 7.2 -> CentOS-2 upgrade. So far it has worked like a charm. The problem is though that yum (no doubt by design) will not downgrade packages to the CentOS versions. The problem is that some old RPMs have 72 in the release (for Red Hat 7.2) and the newer package as 21 in the release (For Enterprise Linux 2.1). Some others are updates to 7.2 which increased the version number where as 2.1 got a backported fix and increased release number. Is there a way to make yum always select a particular Vendor or Distribution? What about using yum to do an explicit downgrade "yum install mypackage-version-oldrelease"? Also, this is an rpm question, does anyone know if I can do "rpm --erase redhat-logos" and "rpm -i centos-logos" at the same time? If I put an obsoletes: tag into centos-logos will yum do that for me? Is that a good idea? At the moment I have a bash script which trys to work out the downgrade rpm commands but I can not check dependencies. Any other suggestions, tips or tricks would be appreciated. John. -- John Newbigin - Computer Systems Officer School of Information Technology Swinburne University of Technology Melbourne, Australia http://www.it.swin.edu.au/staff/jnewbigin

2 2

XF86Setup
by dan1 14 Sep '04

14 Sep '04

Hello. I am looking for the famous XF86Setup command used on RedHat 7.1 (for XFree86) but I didn't find it on CentOS 3.1. Did it change into something else ? Else how can I configure my monitor, mouse, video card settings ? Same questions for starting the hardware material configurator. How can I force it to run ? I know that it runs automatically at boot when we change a card or device, but how to force starting this program manually ? Thanks, Daniel

2 1

I hope this turns out as well as LSB maintiners hope it does
by William Warren 14 Sep '04

14 Sep '04

http://www.eweek.com/article2/0,1759,1645533,00.asp?kc=ewnws091404dtx1k0000… -- My "Foundation" verse: Isa 54:17 No weapon that is formed against thee shall prosper; and every tongue that shall rise against thee in judgment thou shalt condemn. This is the heritage of the servants of the LORD, and their righteousness is of me, saith the LORD. -- carpe ductum -- "Grab the tape"

1 0

up2date-4.2-33-1.centos.3.1
by Michael Rock 14 Sep '04

14 Sep '04

After installing from CD I ran up2date and it looks like I got everything updated. For some reason the up2date icon stays red and tells me that I have up2date-4.2-33-1.centos.3.1 installed but latest version is up2date-4.5.1.centos.3.1 which is the old version of up2date. All the files it wants me to up2date to are the old files. Is this normal for Centos and will it correct itself when new updates come out? thanks all. _______________________________ Do you Yahoo!? Express yourself with Y! Messenger! Free. Download now. http://messenger.yahoo.com

1 0

RHN-like system for CentOS?
by Troy Engel 14 Sep '04

14 Sep '04

Team, Is anyone aware of a Red Hat Network alike system of server package update status that could be adapted for CentOS3? I'm not interested in the complete management aspect (channels, scheduling remote updates, etc) but merely the ability to see your list of systems and determine which are out of date with XXX packages. If you happen to manage a large-ish number of RHES3 systems, you know what a boon this is. ;) Ideas? Thx, -te -- Troy Engel | Systems Engineer Fluid, Inc | http://www.fluid.com

4 3

RE: [Centos] Upgrading Fedora Core 1 to CentOS
by Ho Chaw Ming 14 Sep '04

14 Sep '04

Yes it is all the reason why we want to change to Centos (longer EOL), and we think the stability in RHE should be worth something. It's not always the best to run the newest and the best especially on a production server. So is there anyway to do it other than from yum or to manually force install all the rpm downloaded in the cache? Thanks. regards -----Original Message----- From: centos-admin(a)caosity.org [mailto:centos-admin@caosity.org] On Behalf Of Johnny Hughes Sent: 14 September 2004 19:38 To: CentOS Users Subject: Re: [Centos] Upgrading Fedora Core 1 to CentOS On Tue, 2004-09-14 at 06:24, William Warren wrote: > depends on your definition of downlgrade..i see it as the > oposite..fedora's lifecycle is not nearly as long as RHEL 3..:) Agreed, from a stability and lifecycle perspective, CentOS is much better than (and in my opinion an upgrade to) Fedora Core 1. But from a package perspective, many of the items in an up2date Fedora Core 1 install are newer than the packages in CentOS, so they won't upgrade via yum (since the installed Fedora Core 1 packages are already newer). > > Johnny Hughes wrote: > > > On Mon, 2004-09-13 at 20:51, Johnny Hughes wrote: > > > >>It is a downgrade and not an upgrade to go from CentOS to Fedora > >>Core 1 ... CentOS is an upgrade from RedHat 9 but a Downgrade from FC1. > > > > > > I meant that it is a downgrade to go from FC1 to CentOS ... as RHEL > > 3 was released from RedHat 9.0 code, then FC1 was released later. > > > > Johnny Hughes _______________________________________________ CentOS mailing list CentOS(a)caosity.org http://www.caosity.org/mailman/listinfo/centos

1 0

CentOS 3.1: sshd and pam /etc/security/limits.conf file descriptor settings problem
by Sean Staats 14 Sep '04

14 Sep '04

Why can't non-uid 0 users have more than 1024 file descriptors when logging in via ssh? I'm trying to allow a user to have a hard limit of 8192 file descriptors(system defaults to 1024) via the following setting in /etc/security/limits.conf: jdoe hard nofile 8192 But when jdoe logs in via ssh and does 'ulimit -Hn' he gets '1024' as a response. If he tries to set it with 'ulimit -Hn 8192' he gets an 'Operation not permitted' error. If jdoe instead telnets to the box, he gets the hard limit of 8192 file descriptors. Here is what happens when I set the hard limit to 512 in limits.conf: jdoe hard nofile 512 When jdoe logs in via ssh, he gets a hard limit of 512 file descriptors. The same goes for telnet. So ssh is certainly reading the limits.conf file and applying the settings, so long as nofile <= 1024. Why won't ssh allow users to have more than 1024 file descriptors??? Many thanks! -Sean

1 2

Kickstart and aliased network interfaces
by Ralph Angenendt 14 Sep '04

14 Sep '04

Hi, small(?) problem here: We have several servers which have aliased interfaces (e.g. gigabit: one alias points to the net the backupserver is in, the other one "connects" to the filer the machine needs for work). So I'd like to generate this setup with kickstart (yes, I *could* do that with cfengine or other tools, but want to have the general config in kickstart). Alas: network --bootproto=static --device=eth1:0 --ip=10.65.193.166 --netmask=255.255.255.0 isn't really taken well by anaconda, it barfs up some python-backtrace and stops the installation (it would be neat, if this would be handled friendlier, as in "would you like to continue even though I don't know what you mean"). Any solution for that? Ralph -- Ralph Angenendt......ra(a)br-online.de | .."Text processing has made it possible Bayerischer Rundfunk...HA-Multimedia | ....to right-justify any idea, even one Rundfunkplatz 1........80300 München | .which cannot be justified on any other Tl:089.5900.16023..Fx:089.5900.16240 | ..........grounds." -- J. Finnegan, USC

2 1

← Newer
1
...
5
6
7
8
9
10
11
12
Older →

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Discuss September 2004