Re: [Server-devel] XS 0.5.1 RC - Last round of testing...
On Sun, Feb 15, 2009 at 5:50 AM, Jerry Vonau jvo...@shaw.ca wrote: I've done a back2back comparison between revisor and pungi, given the same kickstart file. Pungi creates media which has the needed openssl.i386 file in it's Packages directory, while revisor creates one that excludes the needed file. Think you found a bug in revisor, I feel that this problem will never get corrected in F9. Interesting find. Going forward, if I recall you were using revisor mainly to embed... Yes, but not the only one: Revisor was supposed to fix a couple of other pungi bugs (skipping xorg deps, and grabbing xs-logos instead of fedora-logos). The Fedora 9 revisor has code that tries to fix those issues and fails. Jeroen is aware of it and said it's fixed in F10. I think he builds a spin (sombrero) that shows that those problems are fixed. The pungi maintainer was also hinting at dropping pungi in favour of revisor. The future is revisor was the between-the-lines msg. So I'm on the fence on this one. Let's get a response from Jeroen. m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] XS 0.5.1 RC - Last round of testing...
On Sat, 2009-02-14 at 15:28 +1300, Martin Langhoff wrote: On Sat, Feb 14, 2009 at 12:31 PM, Jerry Vonau jvo...@shaw.ca wrote: I'm seeing the same errors in the install log as Sameer while installing on an XO. openssl does not get installed, because uname returns as an i586 while there is no openssl.586 in the repo just .686. Just to backup my hunch note that a 586 kernel gets installed as recorded in the install log. yum repolist returns the same error as mentioned in the BZ on the XO. I see Sameer's confirmation. You struck gold with that diagnosis. On the XO OS we have the same problem, and a quick check reveals that we ship openssl i686 on the XS. The disk image build happens on xs-dev which reports i686 surely, but it runs on the XO. Can't find anything in the list archives on the matter. h. m Martin: I've done a back2back comparison between revisor and pungi, given the same kickstart file. Pungi creates media which has the needed openssl.i386 file in it's Packages directory, while revisor creates one that excludes the needed file. Think you found a bug in revisor, I feel that this problem will never get corrected in F9. Going forward, if I recall you were using revisor mainly to embed the kickstart file into the media, given that we run the newly created iso though mkslim anyway,(which can add any files to the iso you want) perhaps we might look at using pungi to spin the release in place of revisor. As a bonus pungi runs a whole pile faster that revisor in creating the installation media. Just my thoughts, Jerry ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] XS 0.5.1 RC - Last round of testing...
On Thu, 2009-02-12 at 23:30 -0800, Sameer Verma wrote: On Wed, Feb 11, 2009 at 10:37 PM, Sameer Verma sve...@sfsu.edu wrote: On Wed, Feb 11, 2009 at 8:42 PM, Martin Langhoff martin.langh...@gmail.com wrote: On Thu, Feb 12, 2009 at 2:41 PM, Sameer Verma sve...@sfsu.edu wrote: was with eth0 not showing up. It looks like I am the victim of the dreaded Realtek 8139 bug. It worked in XS 0.4 but in 0.5.1 it refuses to show up. Strange, but it does look like a driver problem. The links you provide show various different problems with that NIC. In some cases, irqpoll in the kernel boot line fixes, in others some fiddling with ethtool was needed... It'll be good to know which of the fixes helps you :-) appending irqpoll has fixed that problem. Now, I've hit another bug. This is yum broken with _sha256 as stated here. http://fedoraforum.org/forum/showthread.php?t=193507 I'm going to try the workaround. BTW, if you upgraded from XS-0.4, it might be a good idea to rm /etc/udev/rules.d/70-persistent-net.rules No, this was a clean install. I'm running the server for testing only, so I can afford to wipe it clean. Sameer -- Dr. Sameer Verma, Ph.D. Associate Professor of Information Systems San Francisco State University San Francisco CA 94132 USA http://verma.sfsu.edu/ http://opensource.sfsu.edu/ So, after mucking around last night and today, I wiped my XS box and reformatted it to remove ALL traces of 0.4 I have a clean 0.5.1 install on it. md5sum of the ISO is c0fde10b93cab3cb1a3bc3a42ceb5408 I've circumvented the realtek 8139 problem by appending irqpoll in grub.conf That seems to work, although I have to bring up eth0 manually and issuing dhcient eth0 I still hit the bug of _sha256 as mentioned here: https://bugzilla.redhat.com/show_bug.cgi?id=454179 Note that I am not upgrading anything. Its a clean install. I believe the appropriate word for this is: AARGH! I wish Fedora had LTS or Stable branch (it does...kinda...in RHEL...are we allowed to say CentOS here?) but that's another thread and another rant. It does remind me of why I don't run anything on Fedora anymore. Anyway, this is getting in the way. Is anyone seeing this too? If so, then its a significant barrier for 0.5.1 Suggestions? Can you post the /root/install.log and /root/anaconda.log or just send them to me. Jerry ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
