Discuss May 2005

discuss@lists.centos.org

221 participants
344 discussions

CentOS 3 / tg3 driver on an HS20 Blade
by Brendon Colby 20 May '05

20 May '05

Greetings: I know it's been mentioned before that the tg3 driver in CentOS 3.4 is buggy. I had to revert to 3.4 today from 4 (fresh install) and now I have no link and thus no network. I'm wondering if this is caused by this bug and what the work-around is. Thanks. Brendon Colby

3 3

How NOT to have a disk recognized by grub?
by Michael 20 May '05

20 May '05

Greetings: I'm upgrading a fileserver running 3.4 (upgrade to a larger disk). I backed up the data from the "old" disk and slapped in a newer, larger disk and installed Centos-3.4. No problems. Now, there are some files on the "old" disk that I forgot to move to the back-up disk, so I'd like to mount the "old" disk as /dev/hdd and reboot the system and transfer the files [hdd (old disk) --> hda (new disk)]. However, the old disk still has Grub on the MBR and when I boot, the system tries to mount the "/boot" and "/" partitions from BOTH disks! I get errors about duplicate partitions and that those dups won't get mounted.The fileserver does boot but with a configuration combination of both systems. Question: Grub is correctly installed and configured on hda. How do I get the boot process to ignore the old disk (and MBR) on hdd??? I tried google but I can't seem to find this fix. Thanks, Michael Chinn

5 6

Re: [CentOS] Re: mplayer repository for CentOS
by Bryan J. Smith <b.j.smith＠ieee.org> 20 May '05

20 May '05

From: Paul Heinlein <heinlein(a)madboa.com> > FWIW, under RH/FC I typically used Fedora Extras and Livna, grabbing DAG > packages only when I was in a real hurry. I haven't yet tried to > integrate either repo into CentOS 4. I've pretty much just stuck with FE+Lorg for RHEL3Ux/CentOS3.x (EL3) and RHEL4Ux/CentOS4.x (EL4) out of "default," because like yourself, I'm already using them with Fedora Core (I use FC1 for EL3, FC3 for EL4). It's just one of those things where I'm already doing configuration management of my FC systems and integrating in packages from FE (although not MPlayer or most of Lorg, for legal reasons in a corporate environment), so I'm just adopting the same for my EL systems as well. So at home, I do the same (only with more packages from Lorg too). But also like yourself, I have used DAG from time-to-time. -- Bryan J. Smith mailto:b.j.smith@ieee.org

1 0

Re: mplayer repository for CentOS -- oops, send wrong signal ...
by Bryan J. Smith <b.j.smith＠ieee.org> 20 May '05

20 May '05

Maciej Zenczykowski wrote: > is there stuff in livna which isn't in dag/dries > hmm... will have to take a look Preston Crawford wrote: > I haven't found much that I needed that wasn't in dag's repository. > Only two items. Here's what I've found. > gtkpod - for iPods, obviously > xine > mplayer > mplayer-plugin > bittorrent > gkrellm > irssi - ncurses IRC client > mozilla-flash > xmms-mp3 > centericq - ncurses Jabber client > pine, even (yes, I still sometimes use that) > Almost everything I've needed has been there. > And that's a wide range of stuff. Oops, I guess I sent the wrong signal. First off, by far, DAG is going to be the most comprehensive. No way would I dispute that. Secondly, my preference for Fedora Extras/Livna.ORG is because of my experience with Fedora Core and "repository mixing." Nothing against DAG, how he manages to keep up, I have no idea, but I've had better luck with Fedora Extras/Livna.ORG in the past. Now I know many people view it as a catch-22 -- the Fedora Extra team seemed to "keep it itself" in the past, and didn't open up a lot of things. All while Red Hat expected the same level of regression testing on Fedora Extras without putting anyone really on it (and leaving it to the former U of Hawaii team). But as of January of this year, Red Hat has really opened up the formal submission process for including things in Fedora Extras, and put people on regression testing submissions. I'm thinking about submitting several of my own RPMs (e.g., IMAP). Which is why I personally prefer to use Fedora Extras/Livna.ORG by default. I only tap DAG when he clear super-set of packages has something that FE/Lorg doesn't. > The only thing I've installed from downloaded binaries has been Java > stuff, jdk, eclipse, tomcat, etc. And I like to install that stuff > manually anyway. Well, there are some serious licensing issues with some of that (like Java -- far more than "patent" ones), so that is expected. -- Bryan J. Smith mailto:b.j.smith@ieee.org

1 0

Re: iptables port forwarding -- name resolution and not forwarding/routing?
by Bryan J. Smith <b.j.smith＠ieee.org> 20 May '05

20 May '05

From: Johnny Hughes <mailing-lists(a)hughesjr.com> > What I do is have an internal DNS server that does internal IPS for my > domain (it is listed as Primary, no secondaries, for my domain). > Internally, mail.hughesjr.com has the internal address .... externally > it real address. > Internal clients point to the internal DNS server (and internal IP) ... > external clients point to the external IP. Exactomundo. In many cases, it's not a fowarding/routing issue, but a name resolution issue. Private systems are resolving to public addresses, and you want to intercept those from ever reaching a public DNS server. That way you can replace the public name/IP everyone else sees with just the private name/IP that the private LAN should access. That means using private DNS servers and _only_ having the private systems resolving to them. Use "forwarder" and other DNS configuration on the private systems to do external resolution for the internal systems anyway. In addition to solving this issue, you get 2 additional benefits: - DNS cache pooling (all systems are resolving to 2-3 private DNS servers) - UDP/53 restriction (only allow UDP/53 through firewall to those 2-3 private DNS servers) -- Bryan J. Smith mailto:b.j.smith@ieee.org

1 0

Re: [CentOS] RPM dependency graph?
by Bryan J. Smith <b.j.smith＠ieee.org> 20 May '05

20 May '05

From: Christian Nygaard <chris(a)math.uu.se> > Do people know of a tool to generate a visual RPM dependency > graph either in ASCII or with Graphics? Several came out from various Red Hat developers when the "shackles were taken off" and the Red Hat Rawhide team "had at it" with renamed Fedora Core 1 Test in removing years and years of built-up "inter-dependency hell." If you search the [now] Fedora Development lists from August-September 2003, you should see several of them. Some are CLI and some web (PHP?). You can even use them to build pre-assembled package lists for use in Anaconda -- which is really nice if CentOS wants to add its own installation types. E.g., I've always wanted to create a Kerberos Key Distribution Center (KDC aka "key server") install that is rather a minimal package list. -- Bryan J. Smith mailto:b.j.smith@ieee.org

1 0

Iptables - PREROUTING
by David González Romero 20 May '05

20 May '05

I have a server whit 2 interfaces of network, where eth0 is the interfaces connetc to internet and eth1 to the internal network. This server hace a Squid only, but i setting the iptables for protection to the server. Iptables run from script and in this script i setting the redirection for the other server in my internal network to port 80 and 443. I follow the diferent how to and many manual, but the redirect no work, the NAT POSTROUTING work and the squid work to. My DNS is Ok, because the consult whit dig command the answer is the assigned ip. Here put the copy of my script: --------------- Begin the Script #Delete Rules iptables -F INPUT iptables -F OUTPUT iptables -F FORWARD iptables -t nat -F ## Modules install modprobe ip_tables modprobe ip_nat_ftp modprobe ip_conntrack_ftp ## FLUSH of Rules iptables -F iptables -X iptables -Z iptables -t nat -F ## Default policy iptables -P INPUT ACCEPT iptables -P OUTPUT ACCEPT iptables -P FORWARD ACCEPT iptables -t nat -P PREROUTING ACCEPT iptables -t nat -P POSTROUTING ACCEPT ##Filter Begin ##Note: eth0 is the network interface connect to router and ## eth1 comnect to lan ## The localhost full iptables -A INPUT -i lo -j ACCEPT ## DROP the ping for eth0 iptables -A INPUT -i eth0 -p ICMP --icmp-type echo-request -j DROP ## All pack of router acept. iptables -A INPUT -i eth0 -s rango_ip/29 -d 0/0 -p all -j ACCEPT iptables -A INPUT -i eth1 -s 172.16.0.0/24 -d 172.16.0.211/32 -p all -j ACCEPT ##Redirect server web iptables -t nat -A PREROUTING -i eth0 -p tcp --dport 80 -j DNAT --to 172.16.0.3:80 iptables -t nat -A PREROUTING -i eth0 -p tcp --dport 443 -j DNAT --to 172.16.0.3:443 iptables -A FORWARD -i eth1 -p tcp -s 172.16.0.0/24 --dport 80 -j ACCEPT iptables -A FORWARD -i eth1 -p tcp -s 172.16.0.0/24 --dport 443 -j ACCEPT iptables -A FORWARD -i eth1 -p tcp -s 172.16.0.0/24 --dport 53 -j ACCEPT iptables -A FORWARD -i eth1 -p udp -s 172.16.0.0/24 --dport 53 -j ACCEPT ## DNS Full. ACEPT. iptables -A INPUT -p tcp -m tcp --sport 53 -j ACCEPT iptables -A INPUT -p tcp -m tcp --dport 53 -j ACCEPT iptables -A INPUT -p udp -m udp --sport 53 -j ACCEPT iptables -A INPUT -p udp -m udp --dport 53 -j ACCEPT ## SQUID iptables -A INPUT -p tcp -m tcp --sport 3128 -j ACCEPT iptables -A INPUT -p tcp -m tcp --dport 3128 -j ACCEPT iptables -A INPUT -p udp -m udp --sport 3128 -j ACCEPT iptables -A INPUT -p udp -m udp --dport 3128 -j ACCEPT # Close range of port know iptables -A INPUT -s 0.0.0.0/0 -i eth0 -p tcp --dport 1:1024 -j DROP iptables -A INPUT -s 0.0.0.0/0 -i eth0 -p udp --dport 1:1024 -j DROP # Close the port of webmin iptables -A INPUT -s 0.0.0.0/0 -i eth0 -p tcp --dport 10000 -j DROP ## FORWARD # Do log all forward iptables -A INPUT -j LOG --log-level info iptables -A OUTPUT -j LOG --log-level info iptables -A FORWARD -j LOG --log-level info ## Fnish of table FILTER ## General of MASQUERADE iptables -t nat -A POSTROUTING -s 172.16.0.6/32 -o eth0 -j MASQUERADE iptables -t nat -A POSTROUTING -s 172.16.0.10/32 -o eth0 -j MASQUERADE #iptables -t nat -A POSTROUTING -s 172.16.0.1/32 -o eth0 -j MASQUERADE iptables -t nat -A POSTROUTING -s 172.16.0.9/32 -o eth0 -j MASQUERADE ## Do forward to the interfaces of network # Enable of forwarding echo 1 > /proc/sys/net/ipv4/ip_forward /sbin/iptables -L -n > /etc/sysconfig/show /usr/bin/less /etc/sysconfig/show -----------------------------End of Script The output of command "iptables -L -t nat" show me this: Chain PREROUTING (policy ACCEPT) target prot opt source destination DNAT tcp -- anywhere anywhere tcp dpt:www to:172.16.0.3:80 DNAT tcp -- anywhere anywhere tcp dpt:https to:172.16.0.3:443 Some solution?? Regards, David _________________________________________ Tec. David Gonzalez Romero Network/System Administrator CNAP- Centro Nacional Áreas Protegidas Linux counter: 242534 _________________________________________ _________________________________________ Tec. David Gonzalez Romero Network/System Administrator CNAP- Centro Nacional Áreas Protegidas Linux counter: 242534 _________________________________________

2 1

centos-announce spam
by Maciej Zenczykowski 20 May '05

20 May '05

Could we please keep the centos-announce messages in the correct format? Not centos low 3 but low centos 3? Cheers, MaZe

4 3

Re: pronunciation? -- loving CentOS doesn't mean you have to bash Red Hat
by Bryan J. Smith <b.j.smith＠ieee.org> 20 May '05

20 May '05

From: Martyn Drake <martyn(a)drake.org.uk> > To be quite honest with you - that's been and gone so quickly I can't > ever remember what my position was at that time. It actually happened over the span of 2 years before any name change. With the introduction of RHEL as a separate product, RHL was having an identity crisis. It used to be that ".2" was the "enterprise release" (and they even sold SLAs as RHL6.2E), but now there was a 2nd product. Some suits insisted that RHL was still a product -- and some even viewed it as a "competiting" product. I noticed this first hand when RHL7.3, RHL8.0 and RHL9 were "missing" a few things that suddenly "reappeared" later with FC1. Most developers wanted the "shackles off" because RHEL was now the "SLA" product, and RHL was just the incubator for that product now. There was also the continuing nags from CompUSA and other resellers about selling a product that is outdated every 6 months. The result is that Red Hat Linux was deprecated to a project, no longer a product. Vendors largely stopped certifying on RHL after RHL7.3 as the announcements came that updates would only be released for 1 year. Then there were the other projects that added RPMs, etc... and all that finally came to a head in the grand re-design. > I seem to remember it had something to do with Progeny and them > supporting the releases as part of their transition solution. Had _nothing_ to do with Progeny. Progeny just decided to take advantage of the situation and offer $5/month subscriptions in the absence of RHN as of 2003Dec31 (RHL6.2-7.3 EOL) and 2004Apr31 (RHL8.0-9 EOL). Of course, anyone who follows Fedora Legacy knows that even RHL7.3 continued to have updates -- and more ".0" releases like RHL8.0 were dropped due to lack of popularity. > Usually I've gone down the route of using dedicated server providers > that can supply RHEL with the system. That's _exactly_ one of the markets Red Hat was tapping. Red Hat either wants system integrators to use Fedora Core and support their own, or buy RHEL and share the burden with them. > What prompted me to buy my own subscription was that there > were at least two providers that had issues with getting their > RHN satellites/proxies/whatever working and left the machine in > a potentially vunerable state if their install image was a fair a bit > old. That doesn't sound like a Red Hat issue, but a system integrator one. Furthermore, the system integrator _could_ have provided you a YUM repository in the meantime. > Again, I've no problems with this whatsoever. I just wish I could > have purchased a cheaper service with less frills. When Red Hat reaches the economies-of-scale that Microsoft does, then that will happen. But right now, Red Hat has to sell at a price point that is 10x that of Microsoft to get equivalent operating funds for just development. People forget economies of scale is everything. ;-> Most of Red Hat's move into the black came from their Cygnus purchase, and not sales. It wasn't until the subscription model later that they started making a profit atop of their already profitable developer/services unit that Cygnus bought them. > Yggdrasil was my first too. I then worked my way up to Caldera > (spit! spit! spit!) Why? Caldera was a _good_ company until May 2003 (2 months after the original lawsuit, prior to the addedum). Caldera bought SCO and was going to have a split Monterey/IA-64 and Linux/x86 solution. 20 days after the purchase, IBM pulled out of Monterey, yet started to sell their Monterey/Power platform (aka AIX 5L). IBM essentially _squashed_ a _good_ Linux company overnight because it wanted to be the _only_ dual-UNIX/Linux vendor. It saw Monterey/IA-64 (UnixWare 8) as a competitor so it squashed it while it could. Caldera-SCO tried to continue for 18 months, but their entire future was built on low-volume, but high-end sales of UnixWare 8 while their Linux volumes increased. The result was the March 2003 filing, one that I questioned at first, _until_ I read items #50-55. That's when I realized Caldera-SCO was suing for: A) Breach of Contract: Withholding Monterey/IA-64 B) Breach of Non-Compete: Developing Linux/IA-64 Caldera-SCO figured IBM would want the rights to UNIX(R), and IBM would settle in short order. When IBM didn't, they expanded the lawsuit and put up a "smokescreen." They want people to believe their various contractual disputes with IBM, Autozone, Chrysler, etc... are about Linux IP, when they have _no_ ground to stand on. So far, the gambit has worked, largely thanx to the rabid Linux populous who knows nothing about Monetery, and the fact that even the most pro-Linux company is _not_ IBM's friend if you threatened their Power/AIX platforms. Just ask IBM customers who can't buy Linux/x86-64 systems from IBM because IBM tries to sell them Power/AIX, and what IBM does to them if they buy HP Linux/x86-64 systems. ;-> I felt a bit vindicated when Ransom Love, co-founder of Caldera, came out in a Fall 2003 interview with eWeek and basically verified it was all about Monterey, and he didn't agree at all with the May 2003 change of strategy. What really scares me is if and when Caldera-SCO _does_ win on a few counts against IBM, that people will think Caldera-SCO has asserted IP rights on Linux. So far, in all the rulings, on anything Linux, Caldera-SCO has gone no where. But on anything having to do with IBM's Monterey dealings, and possibly "non-compete" violation in the Linux/IA-64 space, the rulings have gone Caldera-SCO's way. I'm afraid a ruling on "non-compete" terms with regard to Linux might be taken as a Linux IP by the media -- and that is very justified given the type of articles we have seen (short of Ransom's interview). To date, Red Hat, Sun and HP have donated about 10x as much in GPL than IBM. People scrutinize Sun for their non-GPL licenses, but don't seem to do the same of IBM -- who has several "non-GPL compatible" licenses on the _majority_ of their so-called "donations." It's like there is the "Linux Quiz Show" and people are tuning in for the money, not the substance. The $1B went to maturing IBM's product line, not Linux donations. The latest $100M is to port 100% proprietary IBM software to Linux, but not open source it. IBM is not our friend. Thank God Linus & co. have watched every donation into the kernel, and made sure no IP was attached (which has held up many of IBM's donations). > but found my footing and went over to Red Hat from that point > onwards. These days I help look after many render farm boxes > and workstations running a combination of Red Hat 7.2 and Fedora Core > 1. We're still converting to FC1 and it's a long and laborious > process given the number of machines that need updating. I adopted APT-RPM for all Red Hat Linux 7+ systems a long time ago. I largely use APT repositories for configuration management. On x86-64, I try to use APT-RPM, but when I need to load both i386 and x86_64 packages, YUM-RPM gets the call. > From that point we'll then going to have to look at 64bit support. YUM-RPM does an outstanding job. My only complaint is that Red Hat stupidly doesn't put the i386 version of Firefox in its repositories alongside the x86_64 packages (it does both for Mozilla though). > Much of our needs are dictated by the software we run, and our > vendors will only support certain distributions. Which is why Studio > Linux http://www.studio-linux.org was born. Why not base it on Fedora Core x86-64? That's one of the things that I'm seeing in Fedora Core like Debian. Instead of just forking from Red Hat Linux never to return (such as Caldera, Mandrake, SuSE, etc... did), people are now forking, but returning to Fedora Core as a base for the next version. That's largely how Debian-based distros work too. > Difficult since we don't have CompUSA here in the UK ;) Or whatever superstores you do have. > The last time I saw a shrink-wrapped Red Hat anything was back in the > days of RH 7, 8 and 9. I even bought some of them to support Red Hat. I have Red Hat Professional Workstation from both RHEL 2.1 and 3. They renamed it (and added volume licenses) the Red Hat Desktop to get the IT media to stop mis-quoting Tiemann that Red Hat _never_ left the desktop. > I'll be good. Has nothing to be with "being good." ;-> I just find that people have an intrinsic feeling that they must define their choice in not only a like of their choice, but a dislike for another. It's not only self-defeating, but many times the facts are not correct. Or my personal favorite, I've seen people state things about Red Hat that actually affect their distro too. I've just seen far too many bugs listed as "Fedora Core-only" when they have been Linux 2.6, Parted, GRUB, cdrecord, etc... issues. -- Bryan J. Smith mailto:b.j.smith@ieee.org

5 5

mplayer repository for CentOS
by Adrian Coman 20 May '05

20 May '05

Hi, Is there a CentOS yum repository for mplayer&dependencies? Thanks, Adrian

2 1

← Newer
1
...
11
12
13
14
15
16
17
...
35
Older →

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Discuss May 2005