Re: Yum and security
On Wed, Jun 15, 2011 at 10:41:14AM -0600, CS DBA wrote: > Hi All; > > We're having discussions within our company about standardizing on a > Linux server OS so we can also standardize our methods & approaches to > configuration, security, etc based on what type of server it is (web, > database, etc). > > That said, we'll need to be able to manage updates in a very controlled > manner. Can anyone point me to any info, documentation, etc in the area > of managing security on RedHat / RPM based distro's, how to manage > security updates via yum, best practices for non yum based packages, etc ? You may also wish to look into how Red Hat Network (Satellite) and its open source upstream, Spacewalk, work. Additionally, you can certainly host your own QA-tested-and-approved yum repos on an https server which you completely control the content thereof, and point your servers at those repos - a lot of people do this. Thanks, Matt -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: bios update
On Sun, May 22, 2011 at 06:21:29AM -0700, James McKenzie wrote: > On 5/22/11 5:51 AM, Craig White wrote: > > and it shouldn't go without mention that Matt is not only a board member > > of Fedora but also has spearheaded tremendous Linux support at Dell > > which provides tools such as these to make life easier for Linux users > > for which I am very grateful. > > > > Craig > > > > > +1. We need more advocates like Matt. Thanks! > [Sidebar] > > Is it still possible to purchase a Dell computer with Linux pre-installed? Yes. All PowerEdge servers and Precision workstations are available with Linux pre-installed, as has been the case for 12+ years. http://dell.com/ubuntu currently lists two notebooks offered in the United States. Other countries may have more or fewer depending on regional sales requirements. Thanks, Matt -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: bios update
On Sun, May 22, 2011 at 05:51:20AM -0700, Craig White wrote: > On Sat, 2011-05-21 at 21:25 -0500, Matt Domsch wrote: > > On Wed, May 18, 2011 at 10:55:24AM -0300, Sebastian wrote: > > >Apologies*if this has been discussed, I couldn't find it. > > >I have a single boot FC14 system on a*Dell precision M6500 precision > > >notebook and*wish to update the BIOS. > > >I go to the Dell download site, and for my machine select the closest > > >match to my operating system: RH Enterprise Linux 5. > > >This gives me the only option to download an EXE file: "M6500A06.EXE" > > >Is it possible to update my bios with this file given my current > > >installation? > > >I have*windows Virtual Box under FC14,*or, > > >Do I need to delete my*installation, install windows and reinstall? > > >As I really don't want to do this, a final option to consider: could I > > >install windows via a USB drive (so as to not destroy my > > > current*system) > > >and then install bios?* > > > > http://linux.dell.com/wiki/index.php/Repository/firmware > > > > describes how to do this for many (but not all) Dell systems. The > > firmware repository on linux.dell.com carries firmware payloads for > > >300 different system types. firmware-addon-dell and firmware-tools > > packages are included in Fedora and EPEL, so you only need the > > firmware payload RPM from the firmware repository then. > > and it shouldn't go without mention that Matt is not only a board member > of Fedora but also has spearheaded tremendous Linux support at Dell > which provides tools such as these to make life easier for Linux users > for which I am very grateful. Former board member (I chose not to stand for re-election in December 2010 due to other commitments, and a strong desire to see others step into that leadership role after my 5 year tenure), but thanks for the votes of confidence! -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: bios update
On Wed, May 18, 2011 at 10:55:24AM -0300, Sebastian wrote: >Apologies*if this has been discussed, I couldn't find it. >I have a single boot FC14 system on a*Dell precision M6500 precision >notebook and*wish to update the BIOS. >I go to the Dell download site, and for my machine select the closest >match to my operating system: RH Enterprise Linux 5. >This gives me the only option to download an EXE file: "M6500A06.EXE" >Is it possible to update my bios with this file given my current >installation? >I have*windows Virtual Box under FC14,*or, >Do I need to delete my*installation, install windows and reinstall? >As I really don't want to do this, a final option to consider: could I >install windows via a USB drive (so as to not destroy my current*system) >and then install bios?* http://linux.dell.com/wiki/index.php/Repository/firmware describes how to do this for many (but not all) Dell systems. The firmware repository on linux.dell.com carries firmware payloads for >300 different system types. firmware-addon-dell and firmware-tools packages are included in Fedora and EPEL, so you only need the firmware payload RPM from the firmware repository then. -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: Help with PPTP VPN connection keeps failing
On Tue, Apr 12, 2011 at 10:21:17PM -0400, Eric B. wrote: > Hi, > > I just recently freshly installed FC14 on my laptop, and have been > trying to get a PPTP VPN connection working, but am having significant > difficulty. > > I've properly configured the VPN settings, and see the connection > succeeds, but then shuts right back down. If I try the same credentials > from within a Windows machine, it connects without fail. [snip] > Apr 12 22:11:55 eric-laptop pppd[3520]: LCP terminated by peer That's the magic message. You need to enable further debugging to figure out why it threw that. http://pptpclient.sourceforge.net/howto-diagnosis.phtml#lcp_term_authentication Thanks, Matt -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: rsync from mirror - problem
On Sun, Apr 17, 2011 at 08:20:11PM -0400, Genes MailLists wrote: > On 04/17/2011 09:33 AM, Genes MailLists wrote: > > > > For the last couple of days or so - my mirror script (I tried a couple > > of mirror sites) seems to keep deleting everything and re-downloading > > each time it runs ... > > > > Mirror script keeps a local copy from which I update all my machines > > here. It runs > > > > rsync -avz --delete --delete-excluded --hard-links > > > > > > Anyone got any ideas why rsync might keep deleting and re-downloading? > (using kernel.org mirror at the moment). No similar reports from others who mirror (including all our public mirrors) so I'm inclined to believe it's a local problem in some way. Thanks, Matt Fedora Mirror Wrangler -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: Slow Mirror: mirror.sr.unh.edu
On Tue, Feb 01, 2011 at 08:32:28PM -0600, Michael Cronenworth wrote: > I was trying to install OpenOffice.org on a fresh F14 machine and > encountered a very slow mirror (mirror.sr.unh.edu). Yum was averaging > 85kb/sec speed. The machine downloading can easily pull 25mbit/sec in a > speed test. Anyone else encountered this? > > The mirror's page says that they are only using 7% of their bandwidth, > so it doesn't seem like they're overloaded. I know Matt sometimes > frequents this list, so if you see this, this is an FYI. I'm getting about 100KBytes/sec from them too, from a few different places. I'll ask their admin. Thanks for the alert. -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: CHECKSUM is not easily accessible on Fedora Download Page
On Fri, Jan 21, 2011 at 01:06:27PM -0500, M?ir?n Duffy wrote: > On Fri, 2011-01-21 at 10:41 +1030, Tim wrote: > > On Thu, 2011-01-20 at 11:22 -0500, M??ir??n Duffy wrote: > > > From talking to numerous novice users in the design of the site I'm > > > not convinced that a checksum file is something that novice users are > > > aware of or much concerned about. > > > > Ignorance is no excuse, as they old saying goes, and it's something that > > needs brought to their attention, with the full how and why. > > > > > The main download link points directly to Fedora's main server, not a > > > mirror, so they'd be downloading the checksum from the same source as > > > the payload anyway. > > > > And the non-main download links...? > > Novice users most likely won't use those. > > > It was always the recommendation, before, to not download from the main > > site, to spread the load around the mirrors. > > Yeh, it was our intention to have mirror manager generate a URL for > those download buttons that made the most sense given geographical > location, but that got dropped due to not having the time. It would be > worth bringing up again. Ah, but it does. :-) The URLs look like: http://download.fedoraproject.org/pub/fedora/linux/releases/14/Live/i686/Fedora-14-i686-Live-Desktop.iso http://download.fedoraproject.org/pub/fedora/linux/releases/14/Fedora/x86_64/iso/Fedora-14-x86_64-DVD.iso which has the goodness of MirrorManager behind it, issuing a HTTP 30x redirect to the best mirror it can. -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: device naming (mentioned in Fedora Weekly News 259)
On Fri, Jan 21, 2011 at 10:44:49AM +0100, Roberto Ragusa wrote: > If there is only one NIC, I want it to be eth0. By default, it will be called em1 or pci#. There is no way to know "there is only one NIC" at any point in time, as each device is discovered in parallel, and devices may be hot-added or hot-removed at any time. To force any one port to be named eth0, you would create an entry for the device in /etc/udev/rules.d/70-persistent-net.rules just as is done today. -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: Fedora x86_64 14 freeze up or reboot immediately, when running a guest in Virtualbox 3.2 or 4.0
On Thu, Jan 06, 2011 at 02:00:00PM +0100, Benjamin Wei?enfels wrote: >Hello, >Fedora x86_64 14 *freeze up or reboot immediately, when running Virtualbox >3.2 or 4.0.*I have installed the repo from Virtualbox. And now it's the >same problem like this guy >had:*[1]http://forums.virtualbox.org/viewtopic.php?f=3&t=36406&start=15 >CPU:*AMD Athlon(tm) 64 X2 Dual-Core Processor TK-53 >Graphics:**ATI Technologies Inc RS690M [Radeon X1200 Series]* Look at this: http://www.virtualbox.org/ticket/7908 I've seen the same problem with openSUSE 11.3 with VB4.0. -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: Is the download.fedora.redhat.com down?
On Fri, Oct 08, 2010 at 01:29:33PM -0400, William Perkins wrote: > Thank you for the suggestion. I have tried the mirrors.kernel.org today > and have found that it will work very well. It is fast and I know that it > will be reliable. I have been very frustrated in my search for new > mirrors to use for Fedora updates. I could not find a new mirror list and > the old one I had or did find had entries for servers that were no longer > in service or had not been updated in years! Bill, I'd appreciate you telling me where you found outdated mirror lists, so I can go remove the stale data and redirect to the "right" lists. http://mirrors.fedoraproject.org has been the canonical listing since Fedora 7. That page is generated by MirrorManager, the tool I wrote that keeps track of all Fedora mirrors, public and private. It currently lists 256 public mirrors, and knows about hundreds of private mirrors too. Thanks, Matt -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: faster /dev/random
On Sat, Aug 21, 2010 at 08:46:24PM -0700, Wolfgang S. Rupprecht wrote: > > Is there an approved way to increase the speed at which the random pool > for /dev/random fills up? I'm playig with dnssec and getnerating 2k rsa > keys is taking up to 3 hours. I've been googling a bit and Intel x86_64 > machines seem to have random number hardware built in (perhaps also > AMD???) Is there a way to funnel this into the entropy pool? Newer Fedora includes rng-utils / rng-tools (it changed names recently) that can use a system's TPM (Trusted Platform Module), which includes a hardware PRNG. rngd can be run as a daemon to get random numbers from the TPM and feed them into the kernel's entropy pool, for use by /dev/random. My Dell servers use this. :-) Thanks, Matt -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: Status of SSDs under Fedora ?
On Fri, Jun 11, 2010 at 10:28:18AM +0100, John Austin wrote: > So the installer does not set "discard" > > A manual > mount -t auto -o discard /dev/sdb2 /mnt/zip > gives > /dev/sdb2 on /mnt/zip type ext4 (rw,discard) > > I assume that setting "discard" in /etc/fstab would work Yes it should. > It would be highly desirable to have definitive answers to > > 1. Is "discard" set by default in F13? No. > 2. How can you find what/where the default mount options are set tune2fs -l /dev/sdb2 though I note that tune2fs in F13 doesn't recognize setting -o discard yet. :-( > 3. Does setting the "discard" option do everything that is required to enable > TRIM to do its job with no further action? > > 4. Is it safe to use the "discard" option yet? > > 5. How do you automatically set the "discard" option when hot-plugging an SSD > disk? >(What controls the mount? hal, udev, XFCE, Gnome, kde, ...) >I am using kdm, XFCE and probably bits of gnome These I don't know. -Matt -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: Appeal to the kmod maintainer of VirtualBox over at rpmfusion
On Wed, Jun 09, 2010 at 03:21:29PM -0500, Michael Cronenworth wrote: > Paul wrote: > > Please, could you rebuild the latest rawhide kernels the modules > > required for VirtualBox-OSE? The latest ones are ages old and won't work > > unless I'm using a seriously old kernel. > > You should post at the rpmfusion lists[1]. > > P.S. this is why I despise kmods. You should convince them to use dkms, > which is supported by Fedora. The akmod/kmod system is just a disaster. (well, as supported as anything to do with 3rd party kernel modules is supported in Fedora). At least it's present and can be useful in these scenarios, and not require someone else to do something for you first. -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: Why are the torrents so slow?
On Tue, May 25, 2010 at 09:53:24AM -0700, Suvayu Ali wrote: > Hi everyone, > > Why are the torrents so slow? I usually get download speeds of up to 600 > Kbps. (both torrent or direct download) But with the 64 bit XFCE torrent > I'm getting downloads @ ~20Kbps. At times the upload spikes to 90Kbps, > but I don't mind that as long as it downloads faster. There are only 7 seeds and 12 leechers on that torrent right now... -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: SSH tunnel for ssh traffic
On Thu, Apr 15, 2010 at 04:12:20PM +0200, Christoph H?ger wrote: > Hi, > > I need to ssh to some remote VM that sit in a private LAN. For any other > service (e.g. RDP) I'd use ssh tunneling just normal. > But what do I do for ssh traffic? Since ssh is not host agnostic, it > will always complain about localhost having a different RSA key. > I just do not want to edit the known_hosts every time I need to connecto > to a new machine! > > Is there some way to tell ssh to use a tunnel directly for a > connection? you want to use ProxyCommand lines in .ssh/config, rather than local redirects. Host deeper-inside HostName 192.168.1.2 ProxyCommand ssh inside nc %h %p Host inside HostName 192.168.0.2 ProxyCommand ssh outside nc %h %p where outside is the public host name/IP, inside is one level inside your private network, directly reachable by host outside, and deeper-inside is 2 levels deep, directly reachable by host inside. $ ssh deeper-inside Thanks, Matt -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: Adding a NIC
On Tue, Apr 13, 2010 at 09:52:25AM -0700, Rick Stevens wrote: > As to how the cards are enumerated...unless the system was installed > with the new NIC already there, it depends on how the mobo scans the > buses. For example, on Dell 2950s, the on-board NICs are seen first, > followed by the PCI cards. On the (more or less equivalent) Dell > 1950s, PCI NICs are seen before the on-board NICs. Go figure. > > Once they're found and /etc/sysconfig/network-scripts/ifcfg-ethX scripts > are created, the HWADDR in those files pretty much keeps things tied > together. If you stuff in a new NIC, it _should_ be the next one in > sequence, regardless which bus it's on as the system generally honors > existing configs (note I said "generally"). A whitepaper I wrote about this problem: http://linux.dell.com/files/whitepapers/nic-enum-whitepaper-v4.pdf http://lkml.org/lkml/2009/3/24/357 http://lwn.net/Articles/356900/ My team is currently trying to get code added to be able to use additional, prettier names, such as "Embedded_NIC_1", to refer to the device names silkscreened on the chassis. The threads above show we've been trying to get better solutions into the kernel and userspace for several years... -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: Another yum update problem
On Tue, Apr 13, 2010 at 12:37:29PM +1000, david walcroft wrote: > On 04/13/2010 04:21 AM, Matt Domsch wrote: > > On Sun, Apr 11, 2010 at 09:41:08AM +1000, david walcroft wrote: > I still get errors. > > updates > | 4.4 kB > 00:00 > http://ga14.files.bigpond.com:4040/fedora/linux/updates/12/i386/repodata/repomd.xml: > > [Errno -1] repomd.xml does not match metalink for updates > Trying other mirror. Thank you for including the http URL above. That helps a lot. That mirror is in fact not up-to-date. The metalink reports this as the expected repodata.xml hash: 9950061445b5cfd8715be2e6ce4b0648f5e219984efd7a36c2391b7ec5c58b7f which corresponds to repomd.xml revision: 1271084436 while the bigpond URL still has this older revision: 1270856692 Now, I admit, there is a bug in MirrorManager which I fixed on Friday but haven't pushed into production yet. In this case, your selected mirror is believed to be up-to-date by MirrorManager, but because the content changed in the last few hours but hasn't yet propogated to this mirror, yum will refuse to use it, knowing that there is a newer version of metadata available. The bug is that MM should account for the propogation delay a little better than it does right now, and allow still fairly recent metadata, even though it's not completely current, to still be used. I plan to get MM fixed in production this week yet, after the beta is out. Thanks, Matt -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: What is in include/linux/kernel.h line 158
On Sat, Apr 10, 2010 at 01:29:08PM -0700, Antonio Olivares wrote: > Dear folks, > > I have encountered this so called sleeping function several times in a good > while (kernel 2.6.27, 2.6.29, 2.6.30, 2.6.31, 2.6.32) Fedora based kernels > and compiled kernels from kernel.org, > > I am asking this because I see this again and again. I would like to know > what is in there so if I can do something about it, or just live with it > since it appears to be harmless, but anyhow it is annoying :( > > BUG: sleeping function called from invalid context at > include/linux/kernel.h:158 > in_atomic(): 0, irqs_disabled(): 1, pid: 9713, name: slmodemd > Pid: 9713, comm: slmodemd Tainted: P 2.6.32.10-90.fc12.i686.PAE #1 > Call Trace: > [] __might_sleep+0xca/0xd1 > [] might_fault+0x19/0x20 > [] copy_to_user+0x34/0x10a > [] amrmo_read+0x50/0x66 [slamr] > [] ? amrmo_read+0x0/0x66 [slamr] > [] vfs_read+0x82/0xe1 > [] ? path_put+0x1a/0x1d > [] sys_read+0x40/0x62 > [] sysenter_do_call+0x12/0x28 The source code to the slamr module (amrmo_read() function) is not present in Fedora or kernel.org. The BUG message is complaining that something (the amrmo_read() function) has disabled interrupts (the irqs_disabled(): 1 part of the message), while at the same time calling a function (copy_to_user()) which can sleep (the process could be unscheduled or mapped out at the moment copy_to_user() is called), so you could wind up deadlocked (interrupts off so nothing to wake the scheduler back up). This is a bug in the slamr driver. Thanks, Matt -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: Another yum update problem
On Sun, Apr 11, 2010 at 09:41:08AM +1000, david walcroft wrote: > I'm using FC12 i386 a fresh install, when trying to update all > programmes or individually I get this error. > > [da...@reddwarf ~]$ sudo yum update yum > Loaded plugins: presto, refresh-packagekit > fedora/metalink > | 4.0 kB > 00:00 > fedora > | 4.2 kB > 00:00 > fedora/primary_db > | 9.7 MB > 00:19 > updates/metalink > | 14 kB > 00:00 > Could not parse metalink > https://mirrors.fedoraproject.org/metalink?repo=updates-released-12&arch=i386 > error was > File /var/cache/yum/i386/12/updates/metalink.xml.tmp is not XML > Error: Cannot retrieve repository metadata (repomd.xml) for repository: > updates. Please verify its path and try again Your link: https://mirrors.fedoraproject.org/metalink?repo=updates-released-12&arch=i386 The "right" link: https://mirrors.fedoraproject.org/metalink?repo=updates-released-f12&arch=i386 spot the missing "f" in yours. :-) -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: How to compile kernel module for a new kernel before reboot?
On Sat, Apr 03, 2010 at 08:21:34PM +0100, Sam Sharpe wrote: > FWIW, DKMS doesn't normally pre-build the module, it builds it on > first boot into the kernel - so it's not exactly what you want. There is now a kernel install-time hook, such that when a new kernel is installed, DKMS can build its modules for that kernel. You don't have to wait for a reboot for the dkms_autoinstaller service to do it then. Thanks for the plugs. -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: FPL steps down: what's the real story?
On Thu, Apr 01, 2010 at 11:23:48PM -0400, Marcel Rieux wrote: >As a matter of fact, as Fedora is mainly financed by Red Hat as a test >bench for RHEL. Fedora is more than a test bench for RHEL. >I can hardly see how Fedora could stand as "a completely >separate entity". Legally, Fedora (the trademark) is owned by Red Hat, and the Fedora Project Board is given the authority to manage it. Red Hat is also the largest (but by no means the only) Fedora sponsor. http://fedoraproject.org/sponsors The Fedora EMEA group does stand as a completely separate entity, with its own governance, membership, income, expenses, etc. In the US however, that isn't as feasible for a not-for-profit with one primary sponsor, so we didn't do that. >CentOS and Scientific Linux are separate entities from >Red Hat, not Fedora. > >Maybe this should be made clearer so that developers understand what kind >of project they're involved in. There are advantages working for a major >Linux distribution such as Red Hat. Are there enough, I don't know. This >is a question I raise in the case study I'm about to submit. Unless you receive a paycheck directly from Red Hat as an employee, you are not "working for a major Linux distribution" when you contribute to Fedora. You are contributing to Fedora, or one of the upstream projects that Fedora draws from. Red Hat may be one additional beneficiary of that work, but the Red Hat productized software starts as a subset of the Fedora software, and then adjusted to best suit Red Hat's customers (in Red Hat's opinion). Nothing is stopping someone else from doing likewise - which, as you note, CentOS and SL do similarly. The Fedora Project Board is comprised of 4 appointed, 5 community-elected seats, and the chair (FPL) is by definition a Red Hat employee. >It's not time to discuss this here but I certainly believe that >developers' contribution should be more fully acknowledged, and I mean >this not only in an abstract manner. For the unrest to cease -- because >there is some unrest -- the relation between development and management >will have to evolve, just to make sure that it's impossible from now on >for a CEO and his wife to run away with hundreds of millions $, leaving >developers sixpence none the richer(1). > >(1) Of course, this is now impossible, but a sense of balance must still >be established. What are you proposing, exactly? >When you ask developers to work, at least at the beginning, for free, you >must play an honest game. Otherwise, you won't get the best. There should >be a dynamic way to define when the beginning is being stretched too far, >without tying development and management by any obligation. > >Investors also will gain from a development model that works. I'm confused. If you want Red Hat to treat you as an employee, that seems to be here: http://www.redhat.com/about/careers/ If you want to create your own commercial Linux distribution, using Fedora as your upstream, you're welcome to do that too. It's easy to create a new distro (just look at the lwn.net list of distros page). It's much harder to sustain one, build a following, and figure out a business model that will be self sufficient. A few companies have done so. And if you have other goals, and by participating in Fedora you can achieve them, fantastic. -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: Looking for a monitor driver
On Tue, Mar 30, 2010 at 03:25:48PM -0400, Margaret Doll wrote: > We are trying to attach a Dell E2210H monitor to a Dell X260 Optiplex > which is running 2.6.10-1.771_FC2. > > I have found the drivers for Windows but not for Fedora. Where can I > access monitor drivers for Fedora? As noted in this thread, there are no separate "drivers" to install for this monitor. I did update the hwdata MonitorsDB database file in August 2009 to include these lines: Dell; Dell E2210H(Analog); DELD030; 30.0-83.0; 50.0-76.0; 1 Dell; Dell E2210H(Digital); DELD031; 30.0-83.0; 50.0-76.0; 1 On FC2 I couldn't begin to tell you where the MonitorsDB file is though... Thanks, Matt P.S.: I do periodically grab all the Windows *.inf files for all Dell monitors, as published on support.dell.com, and update the MonitorsDB file in the hwdata package with them. On any newish Linux version though, this data is extracted from talking to the monitor directly, rather than looked up in the MonitorsDB file. The Windows .inf format for monitors is fairly limited in terms of describing the various modes, and in particular, native resolution, of a given LCD; the Linux MonitorsDB format even more so. Much better to use a newer version of Xorg that can do the ECC directly. -- Matt Domsch Technology Strategist Dell | Office of the CTO -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Looking out for stale mirrors
FYI. On Monday 3/22, I changed how Fedora's MirrorManager treats "stale" mirrors (mirrors with valid, but not quite up-to-date, content). Previously, mirrors could be at most 7 days stale, and still be offered via the mirrorlists, to yum. Now, that is at most 2 days stale. This allows for sufficient propogation delay for new updates, while ensuring that our users are getting the latest updates that are available. If you see yum frequently falling back to use another mirror, please let mirror-ad...@fedoraproject.org know which mirror you see it falling back from (the "stale" mirror), so we can investigate with the mirror owner. Thanks, Matt Fedora Mirror Wrangler -- Matt Domsch Technology Strategist, Dell Office of the CTO linux.dell.com & www.dell.com/linux -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: From when on is IPv6 address information available on an interface?
On Mon, Mar 15, 2010 at 02:20:55PM +0100, fedora wrote: > As you can see from the DNS update command list, the IPv4 address is > present, whereas the IPv6 address is empty. The IPv6 address is asked > for via the command > ip -o -f inet6 addr show dev wlan0 scope site > which returns an empty string. See MinRtrAdvInterval and MaxRtrAdvInterval in your /etc/radvd.conf on your router that advertises the IPv6 prefix. Default is minimum 30 seconds, max 100 seconds, which in means you're slipping in to bring up the interface and set your DDNS between announcements. Thanks, Matt -- Matt Domsch Technology Strategist, Dell Office of the CTO linux.dell.com & www.dell.com/linux -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: [Request] Adding the DebugInfo packages to the Mirrors
On Sun, Feb 21, 2010 at 12:30:26AM +0530, Sawrub wrote: >FTP listing of the Fedora mirrors hosted in India, is missing the >debuginfo packages. I'm able to locate the same under most of the >International mirrors with few of them to list picked from the mirror list >hosted at [1]http://mirrors.fedoraproject.org/publiclist/Fedora/12/ are As you note later, yum will eventually fall back to finding a mirror somewhere globally, though perhaps not locally, that contains the debuginfo packages. So while it may not be quite as fast for you to download them, they are available. Our mirrors do a fantastic job for Fedora - we couldn't handle the several million installed systems worldwide without them. At the same time, each mirror has its own constraints on what resources they can provide. A full Fedora mirror alone takes just under 1TB right now, which is quite a lot to keep on every mirror. Mirrors therefore can be selective about what they carry, the only requirement being that they exclude at a whole directory level (they can't mirror only some files in any single directory). Historically, debuginfo hasn't been all that useful. With abrt being more widely used, and people being prodded through the GUI to download debuginfo to submit better bug reports, debuginfo is being used quite a bit more than before. But it's still a small number of downloads, but fairly large in disk space size, in comparison to all the other packages in the distro. So, a mirror, trying to be efficient in the use of their disk space and download bandwidth, often will exclude the debuginfo packages exactly because they are large and infrequently requested. I'll send a note to the mirrors asking them to consider carrying debuginfo, but if they don't wish to, I can't make them, nor would I even try. Thanks, Matt Fedora Mirror Wrangler -- Matt Domsch Technology Strategist, Dell Office of the CTO linux.dell.com & www.dell.com/linux -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
Re: Dell R710 Server and FC11
On Thu, Jan 28, 2010 at 08:57:17PM +0800, Edward S.P. Leong wrote: > Dear All, > > Would you mind to tell me does the DELL R710 Server and Raid Controller > support with FC11 OS ? > http://www1.ap.dell.com/hk/en/business/servers/server-poweredge-r710/pd.aspx?refid=server-poweredge-r710&s=bsd&cs=hkbsd1 Yes. Not that Dell officially supports such, but it works. Tools such as Dell OpenManage are not expected to work on Fedora. You just can't call Dell Technical Support with OS-related questions (you of course can call regarding hardware issues). You may also wish to join the linux-poweredge mailing list (http://lists.us.dell.com/mailman/listinfo/linux-poweredge) which is for system administrators of Dell PowerEdge servers. Thanks, Matt -- Matt Domsch Technology Strategist, Dell Office of the CTO linux.dell.com & www.dell.com/linux -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Yum update crashes
On Fri, Jan 29, 2010 at 01:27:30PM -0500, Jim wrote: > FC12-X86_64/KDE > > Yum update is crashing. > I did a yum clean all, but that doesn't help. > > > ]# yum update > Loaded plugins: presto, refresh-packagekit > fedora/primary_db > > > | 12 MB 00:39 > google-chrome > > > | 951 B 00:00 > google-chrome/primary > > > | 1.8 kB 00:00 > google-chrome > > > 3/3 > livna > > > | 2.4 kB 00:00 > livna/primary_db > > > | 2.7 kB 00:00 > rpmfusion-free > > > | 3.8 kB 00:00 > rpmfusion-free/primary_db > > > | 344 kB 00:05 > rpmfusion-free-updates > > > | 3.3 kB 00:00 > rpmfusion-free-updates/primary_db > > > | 159 kB 00:00 > rpmfusion-nonfree > > > | 3.3 kB 00:00 > rpmfusion-nonfree/primary_db > > > | 91 kB 00:00 > rpmfusion-nonfree-updates > > > | 3.3 kB 00:00 > rpmfusion-nonfree-updates/primary_db > > > | 40 kB 00:00 > updates/metalink > > > | 14 kB 00:00 > updates > > > | 4.4 kB 00:00 > updates/primary_db > > > | 3.7 MB 00:16 > Error: file is encrypted or is not a database This could be your yum history database getting corrupted. $ sudo rm -f /var/lib/yum/history/* (This will of course keep you from rolling back transactions from your history, as it'll be gone.) -- Matt Domsch Technology Strategist, Dell Office of the CTO linux.dell.com & www.dell.com/linux -- users mailing list users@lists.fedoraproject.org To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: Database disk image is malformed
On Fri, Jan 22, 2010 at 09:35:49PM -0500, Jim wrote: > FC12-x86_64/KDE > > #yum install > > Error: Database disk image is malformed > > I removed /var/lib/rpm/__db.** files > And then I did a ; > #rpm --rebuilddb * > > But that still doesn't help when running yum. > What am I not doing right? I ran into this once recently, when I hit Ctrl-C at an inopportune moment while yum was running on that window. Like you, thought it was the RPM database that was corrupt, and like you did the same thing to no effect. # yum clean all had no effect either. To resolve, I deleted all the contents of /var/lib/yum/. It's kind of heavyweight, but it resolved it. You may try simply deleting /var/lib/yum/history/* first, and see if that's sufficient (assuming you don't need to be able to have yum roll back completed transactions to some previous state). If not, try deleting yumdb/ there. There may be a better way to handle this, but I'll let the yum experts chime in then. Thanks, Matt -- Matt Domsch Technology Strategist, Dell Office of the CTO linux.dell.com & www.dell.com/linux -- users mailing list users@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
Re: possible bad ipv6 mirror [partial solution]
On Thu, Jan 14, 2010 at 10:40:23PM -0600, Michael Cronenworth wrote: There was a question about if IPv6 GeoIP in MirrorManager was working. Yes, it is. >On 01/14/2010 06:20 PM, Michael Cronenworth wrote: > > It's tiresome that yum does not detect bad URLs. ?? Please advise exactly what URL you're getting from the mirror list that is "bad". There are no URLs with '[' being returned as in an IPv6 address, they're all returned by host name, and yum resolves that into either an IPv6 or IPv4 address. >Added mirror.switch.ch to point to ::1 to my hosts. Yum is now using >[2001:400:201:6::35] to pull updates. Working great. > >Hmm I'm using a tunnel from sixxs.net at home to test this. Using a mirrorlist.txt file with entry: http://[2001:620:0:8::20]/ftp/mirror/fedora/linux/releases/12/Everything/x86_64/os/ yum works fine for me. Likewise, using a mirrorlist.txt file with entry: http://mirror.switch.ch/ftp/mirror/fedora/linux/releases/12/Everything/x86_64/os/ yum works fine with this, using the IPv6 address. $ traceroute6 mirror.switch.ch traceroute to mirror.switch.ch (2001:620:0:8::20), 30 hops max, 80 byte packets 1 my-home-ipv6-gateway () 1.587 ms 1.515 ms 1.811 ms 2 gw-6.qas-01.us.sixxs.net (2001:4830:1600:5::1) 57.235 ms 57.310 ms 57.528 ms 3 sixxs-gw.hotnic.us.occaid.net (2001:4830:e6:7::1) 57.643 ms 57.788 ms 57.911 ms 4 10gigabitethernet2-2.core1.ash1.he.net (2001:504:0:2::6939:1) 63.074 ms 63.263 ms 65.320 ms 5 10gigabitethernet1-2.core1.nyc4.he.net (2001:470:0:36::2) 70.825 ms 70.949 ms 71.085 ms 6 10gigabitethernet1-2.core1.lon1.he.net (2001:470:0:3e::2) 137.648 ms 132.130 ms 132.103 ms 7 10gigabitethernet1-1.core1.ams1.he.net (2001:470:0:3f::2) 139.992 ms 140.341 ms 140.472 ms 8 10gigabitethernet1-1.core1.fra1.he.net (2001:470:0:47::2) 145.489 ms 145.496 ms 145.610 ms 9 10gigabitethernet1-1.core1.zrh1.he.net (2001:470:0:10d::2) 150.030 ms 150.086 ms 153.936 ms 10 swiIX2-10GE-3-2.switch.ch (2001:7f8:24::35) 153.997 ms 154.223 ms 154.280 ms 11 swiEZ2-10GE-1-3.switch.ch (2001:620:0:c00a::1) 154.526 ms 154.785 ms 155.077 ms 12 swiCS3-10GE-1-1.switch.ch (2001:620:0:c02e::1) 155.099 ms 155.245 ms 155.240 ms 13 mimas-nxge0.switch.ch (2001:620:0:8::20) 155.369 ms 152.898 ms 152.965 ms Is it possible that there's some IPv6 breakage between you and that site? Thanks, Matt Fedora Mirror Wrangler and MirrorManager author -- Matt Domsch Technology Strategist, Dell Office of the CTO linux.dell.com & www.dell.com/linux -- users mailing list users@lists.fedoraproject.org To unsubscribe: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
