Re: The joys of rawhide returning to form
On Wed, 2006-03-22 at 09:54 -0800, darrell pfeifer wrote: > For the last couple of weeks as rawhide settled into FC5 I had a > completely working system. Now that the freeze has been lifted for a > couple of days I yum updated this morning. I'm seeing all of the same things. > Printing broke. > > There was a cups update. I think I saw a scriptlet fail. There appears > to be no default printer. Using the control panel to remove the > current (USB) printer and remake a new one doesn't help. Test printing > fails. > > X broke > > There was an X server update. I have an ATI Radeon R300 so am involved > in the DRI fray. In previous versions commenting out the load of the > DRI module worked but in the new version DRI looks like it is still > attempted and I get the black screen of death. Reverted to the X > server rpm off the FC5 cd. In my case I had the EXA acceleration enabled on X, and that appeared not to be loading. I was able to get X working again by reconfiguring. > ipw2200 broke > > There was a kernel upgrade (2074) and now the driver won't load. > Booted the 2064 kernel which continues to work. > > So that was no X, no wireless and no printing. Eeeks. Thank goodness I have a wired jack near to where I use my laptop at home... Jeff signature.asc Description: This is a digitally signed message part
Re: Anaconda: good work!
On Wed, 2006-03-22 at 11:45 -0500, Jeremy Katz wrote: > On Wed, 2006-03-22 at 10:12 -0500, Dimi Paun wrote: > > First impression: anaconda really rocks! This is a > > very solid, pleasant installer. > > Thanks -- always nice to see people not just flaming us :-P *snip* I have to agree that the new installer is very nice. But despite this I really wish you will apply some bugfixes to anaconda and release it in updates. This of course does not help with FC5, but it will help those who use anaconda to make new distros or respins such as I plan to do (either official or unofficial). The complete lack of updates to anaconda in FC4 bugged me, but I was too late in the cycle to make you reconsider. This time around I hope you will atleast take obvious bugfixes and maintain the FC5 anaconda atleast for a little while. I'm not asking for new features, only fixes. -HK
Intelligent Anaconda (was: Re: The Morning After)
On Thu, 2006-03-23 at 01:45 +0530, Rahul Sundaram wrote: > The fact that we can do a minimal installation with a > single CD, a desktop class installation in 2 CD's or a network > installation with boot.iso image is far from clear to everyone involved. > It has been offered for years and people continue to request it as a new > feature not being aware that it exists already in just about every > release. Kickstart should be much better advertised. In my opinion when you boot anaconda from CD1, it should ask you what cds you have available. So if I only downloaded CD1 & 2 it would not allow me to select packages that are not available on those two cds, possibly asking me whether I want yum to download what I don't have. This selection dialog could also be extended to let you chose that you have indeed downloaded extras CD1, CD2 and CD3 (or somesuch). Or let you select "Other CD" and let anaconda scan the disk and find the prepared-by-anaconda list of rpms and deps on the disk. This would allow me to make my own addons cd that would just be installed from just like the other cds. Yes, yum has solved this problem partially but not everyone has 24/7 highspeed Internet access. And some just don't want to have to install from the internet every time. -HK
Re: Unable to mount cd & dvd after upgrading to FC5
Jeff Spaleta wrote: I solved the problem, I guess. Here's what I've done: I've installed hal-gnome package, turned on hal service on runlevel 5, commended out some fstab lines and rebooted machine. Well, gnome-mount crashes often, and I got some growfs error message after I successfully burned multisession DVD (?), which was ok thoe, but at least I can say that it works for now. After these unpleasent experiences, I can't say I'm gonna recommend upgrading from FC4 to FC5 instead of clean install. And here's my fstab, so you can see what lines I commented out, and what lines I took from my test system and manually added: [ija...@munja ~]$ cat /etc/fstab LABEL=/1/ ext3defaults1 1 LABEL=/boot1/boot ext2defaults1 2 #none/dev/ptsdevpts gid=5,mode=620 0 0 #none/dev/shmtmpfs defaults0 0 LABEL=/home /home xfs defaults1 2 #none/proc procdefaults0 0 #none/syssysfs defaults0 0 #LABEL=SWAP-hda5 swapswapdefaults0 0 devpts /dev/ptsdevpts gid=5,mode=620 0 0 tmpfs /dev/shmtmpfs defaults0 0 proc/proc procdefaults0 0 sysfs /syssysfs defaults0 0 /dev/hda5 swapswapdefaults0 0 > On 3/22/06, Igor Jagec wrote: >> I even tried to run hal deamon manually, to play >> with gnome-mount and so on, but all of that didn't help. I tried to 'rpm >> -V hal', but I got no output. Is there any way to solve that problem >> manually? To make hal to detect my hdc and hdd devices? Any help would >> be highly appreciated. > 1) if you tried to run it manually... does that mean it wasn't running > already? Most likely it wasn't. > 2)is the dbus stuff running correctly? /sbin/service messagebus status [r...@munja ~]# /sbin/service messagebus status dbus-daemon (pid 4147 1558) se izvršava... Which means it runs properly. BTW I tried to get english output with 'export LANG=en_EN.ISO8859-1', and it didn't help for that command, and for some it did (?). Never mind. > 3) does the outout of lshal show your hdc and hdd devices? [r...@munja ~]# lshal|grep hdc block.device = '/dev/hdc' (string) linux.sysfs_path_device = '/sys/block/hdc' (string) linux.sysfs_path = '/sys/block/hdc' (string) [r...@munja ~]# lshal|grep hdd storage.cdrom.hddvdrw = false (bool) storage.cdrom.hddvdr = false (bool) storage.cdrom.hddvd = false (bool) storage.cdrom.hddvdrw = false (bool) storage.cdrom.hddvdr = false (bool) storage.cdrom.hddvd = false (bool) block.device = '/dev/hdd' (string) linux.sysfs_path_device = '/sys/block/hdd' (string) linux.sysfs_path = '/sys/block/hdd' (string) block.device = '/dev/hdd' (string) linux.sysfs_path_device = '/sys/block/hdd/fakevolume' (string) linux.sysfs_path = '/sys/block/hdd/fakevolume' (string) > there should be a block of output starting with udi = something > and ending with linux.sysfs_path = something > for both the hdc and hdd device That above is an output after I solved the problem. Since gnome-mount crashes often, I'm not quite sure I solved the problem completely, but at least it works now. > 4) I'm still not sure what you attempted exactly, so its pretty > difficult to provide any feedback. I didn't know how to provide you more information, but I hope that above will help a bit. I saw on the redhat's news group that I'm not the only one who experienced that problem. -- Igor Jagec
Re: Anaconda: good work!
Hi. On Wed, 22 Mar 2006 12:36:33 -0500, Ray Strode wrote: > I think installing gconf schemas got slower with the gconf backend > changes. This may have something to do with things, not sure. > > If that is a significant cause of the slowdowns, we can partially > alleviate the problem, by changing all %post snippets from > > for S in ""; do > gconftool-2 --make-install-rule $S > done > > to > > gconftool-2 --make-install-rule This step (what does that do, anyway?) is _slow_. 30 seconds per schema on my 500MHz iBook. Updating gnome-games takes ages.
strange X11 Problems with cisco vpn client under Fedora 4 or 5
Hi, first of all, sry when this is not the right mailing list. I have problems with the cisco vpnclient 4.xxx under Fedora 4 or 5. The client works perfectly, but when it is started and i login with a user into X11, after a while the mouse pointer stacks at the left side. This happens ONLY when the cisco vpn client is connected with my office. So my question is, how can this happen, where can i look or debug to get more info? I have to reboot the computer, deleting modules restarting X do not work. Anyone here with the same Problem? On my notebook with Fedora Core 3 i do not have this problem. Thanks for reading, Joern Rink
Re: Anaconda: good work!
Igor Jagec wrote: Paul Howarth wrote: I upgraded a desktop box from FC4->FC5 yesterday, and my perception was Can you mount CD or DVD medias? I've upgrade FC4 via install media. Everything works fine on FC5 clean install, but not on upgrade from FC4. Didn't try; I always do NFS installs as they're so much faster than DVD/CD. Paul.
Re: Anaconda: good work!
On Wed, 2006-03-22 at 11:45 -0500, Jeremy Katz wrote: > On Wed, 2006-03-22 at 10:12 -0500, Dimi Paun wrote: [snip] > One thing that we really probably need to look at for FC6 > is revisiting the time remaining algorithm. Yeah, it took about fifteen minutes after all packages were installed to finish doing an update. That was long "1 minute remaining". :-) -- Garry Williams -- +1 (678) 656-4579
Re: The joys of rawhide returning to form
On 3/22/06, darrell pfeifer wrote: > So that was no X, no wireless and no printing. Eeeks. -jef" Breaking News! Rawhide eats babies!!! Film at 11 "spaleta
Re: Anaconda: good work!
On 3/22/06, Thomas Canniot wrote: > Maybe we could do something here that really helps people, newbies who > are just installing their first linux distribution. I dreamt of an > anaconda that helps newbies make their first steps in Fedora Core. Screw that... embedded game of nexuiz would be much better. As for the mock up... perhaps an embedded ogg video of desktop interactions with localized closed caption text ala annodex. -jef"so what if anaconda would require a minimum of 2 gig of ram for the video to play while anaconda does packaging actions."spaleta
Re: FC5 kickstart i386 and x86_64
> I've downloaded FC5 for i386 and x86_64. I do netboot kickstart > installs. With FC3 and FC4, and previous I think, If I didn't have a > network stanza in the kickstart config file it would just bypass > configuring the interfaces and just use the one that was selected during > the boot process. With FC5 it asks to configure all the interfaces. If > I add a single network line to the config file it bypasses the interface > configuration again and configures on the interface in the config file. > My machines don't all use the same interface for network connectivity, > and I liked the way it worked before. Any suggestions would be > appreciated. Yes I have RTMF. Kickstart got a major overhaul for FC5, so it's likely I have accidentally dropped some behavior that was present in earlier versions. In general, the way we do things is that kickstart asks for any missing information so in this case, asking for the networking is consistent with that policy. However if it's a regression from earlier versions, file a bug against anaconda and I'll take a look at it. - Chris
gnome-screen-saver unlock artwork
I have some inconsistencies across several machines with the artwork for the gnome screen saver unlock window that I thought was due to different initial fc5 test installs. Now that I've done clean installs of fc5 proper on the same systems the inconsistencies remain. On some systems I have the themed fedora core 5 unlock window and on others I have a simple grey box. I can't really do a screen shot because the screen is locked. It doesn't appear to be user related because a new user on an affected system still has just the grey box unlock dialog. Has anyone else seen this? tjb -- === | Thomas Baker email: t...@unh.edu| | Systems Programmer | | Research Computing Center voice: (603) 862-4490 | | University of New Hampshire fax: (603) 862-1761 | | 332 Morse Hall | | Durham, NH 03824 USA http://wintermute.sr.unh.edu/~tjb | ===
rawhide report: 20060323 changes
Updated Packages: (none) Broken deps for s390x -- avalon-logkit - 1.2-3jpp_3fc.noarch requires servletapi5 castor-demo - 0.9.5-1jpp_2fc.noarch requires servletapi5 geronimo-specs - 1.0-0.M2.2jpp_7fc.s390x requires servletapi5 hsqldb - 1.80.1-1jpp_8fc.s390x requires servletapi5 jakarta-commons-fileupload - 1:1.0-3jpp_5fc.noarch requires servletapi5 jakarta-taglibs-standard - 1.1.1-4jpp_3fc.noarch requires servletapi5 >= 0:5.0.16 rhythmbox - 0.8.8-2.s390x requires libgstcontrol-0.8.so.1()(64bit) rhythmbox - 0.8.8-2.s390x requires libgstgconf-0.8.so.0()(64bit) rhythmbox - 0.8.8-2.s390x requires libgstreamer-0.8.so.1()(64bit) struts - 1.2.8-2jpp_9fc.s390x requires servletapi5 struts-webapps-tomcat5 - 1.2.8-2jpp_9fc.s390x requires tomcat5 velocity - 1.4-3jpp_4fc.noarch requires servletapi5 xalan-j2-demo - 2.6.0-3jpp_9fc.s390x requires servletapi5 Broken deps for i386 -- GFS-kernel - 2.6.15.1-5.FC5.17.i686 requires /lib/modules/2.6.15-1.2054_FC5 GFS-kernel - 2.6.15.1-5.FC5.17.i686 requires kernel = 0:2.6.15-1.2054_FC5 GFS-kernel-smp - 2.6.15.1-5.FC5.17.i686 requires /lib/modules/2.6.15-1.2054_FC5smp GFS-kernel-smp - 2.6.15.1-5.FC5.17.i686 requires kernel-smp = 0:2.6.15-1.2054_FC5 GFS-kernel-xen0 - 2.6.15.1-5.FC5.17.i686 requires kernel-xen0 = 0:2.6.15-1.2054_FC5 GFS-kernel-xenU - 2.6.15.1-5.FC5.17.i686 requires /lib/modules/2.6.15-1.2054_FC5xenU GFS-kernel-xenU - 2.6.15.1-5.FC5.17.i686 requires kernel-xenU = 0:2.6.15-1.2054_FC5 cman-kernel - 2.6.15.1-0.FC5.16.i686 requires /lib/modules/2.6.15-1.2054_FC5 cman-kernel - 2.6.15.1-0.FC5.16.i686 requires kernel = 0:2.6.15-1.2054_FC5 cman-kernel-smp - 2.6.15.1-0.FC5.16.i686 requires /lib/modules/2.6.15-1.2054_FC5smp cman-kernel-smp - 2.6.15.1-0.FC5.16.i686 requires kernel-smp = 0:2.6.15-1.2054_FC5 cman-kernel-xen0 - 2.6.15.1-0.FC5.16.i686 requires kernel-xen0 = 0:2.6.15-1.2054_FC5 cman-kernel-xenU - 2.6.15.1-0.FC5.16.i686 requires /lib/modules/2.6.15-1.2054_FC5xenU cman-kernel-xenU - 2.6.15.1-0.FC5.16.i686 requires kernel-xenU = 0:2.6.15-1.2054_FC5 dlm-kernel - 2.6.15.1-0.FC5.14.i686 requires /lib/modules/2.6.15-1.2054_FC5 dlm-kernel - 2.6.15.1-0.FC5.14.i686 requires kernel = 0:2.6.15-1.2054_FC5 dlm-kernel-smp - 2.6.15.1-0.FC5.14.i686 requires /lib/modules/2.6.15-1.2054_FC5smp dlm-kernel-smp - 2.6.15.1-0.FC5.14.i686 requires kernel-smp = 0:2.6.15-1.2054_FC5 dlm-kernel-xen0 - 2.6.15.1-0.FC5.14.i686 requires kernel-xen0 = 0:2.6.15-1.2054_FC5 dlm-kernel-xenU - 2.6.15.1-0.FC5.14.i686 requires /lib/modules/2.6.15-1.2054_FC5xenU dlm-kernel-xenU - 2.6.15.1-0.FC5.14.i686 requires kernel-xenU = 0:2.6.15-1.2054_FC5 gnbd-kernel - 2.6.15-5.FC5.23.i686 requires /lib/modules/2.6.15-1.2054_FC5 gnbd-kernel - 2.6.15-5.FC5.23.i686 requires kernel = 0:2.6.15-1.2054_FC5 gnbd-kernel-smp - 2.6.15-5.FC5.23.i686 requires /lib/modules/2.6.15-1.2054_FC5smp gnbd-kernel-smp - 2.6.15-5.FC5.23.i686 requires kernel-smp = 0:2.6.15-1.2054_FC5 gnbd-kernel-xen0 - 2.6.15-5.FC5.23.i686 requires kernel-xen0 = 0:2.6.15-1.2054_FC5 gnbd-kernel-xenU - 2.6.15-5.FC5.23.i686 requires /lib/modules/2.6.15-1.2054_FC5xenU gnbd-kernel-xenU - 2.6.15-5.FC5.23.i686 requires kernel-xenU = 0:2.6.15-1.2054_FC5 Broken deps for ia64 -- rgmanager - 1.9.31-3.ia64 requires ccs Broken deps for ppc64 -- avalon-logkit - 1.2-3jpp_3fc.noarch requires servletapi5 castor-demo - 0.9.5-1jpp_2fc.noarch requires servletapi5 emacs - 21.4-5.ppc64 requires fonts-xorg-75dpi geronimo-specs - 1.0-0.M2.2jpp_7fc.ppc64 requires servletapi5 hsqldb - 1.80.1-1jpp_8fc.ppc64 requires servletapi5 jakarta-commons-fileupload - 1:1.0-3jpp_5fc.noarch requires servletapi5 jakarta-taglibs-standard - 1.1.1-4jpp_3fc.noarch requires servletapi5 >= 0:5.0.16 struts - 1.2.8-2jpp_9fc.ppc64 requires servletapi5 struts-webapps-tomcat5 - 1.2.8-2jpp_9fc.ppc64 requires tomcat5 velocity - 1.4-3jpp_4fc.noarch requires servletapi5 xalan-j2-demo - 2.6.0-3jpp_9fc.ppc64 requires servletapi5 Broken deps for s390 -- avalon-logkit - 1.2-3jpp_3fc.noarch requires servletapi5 castor-demo - 0.9.5-1jpp_2fc.noarch requires servletapi5 geronimo-specs - 1.0-0.M2.2jpp_7fc.s390 requires servletapi5 hsqldb - 1.80.1-1jpp_8fc.s390 requires servletapi5 jakarta-commons-fileupload - 1:1.0-3jpp_5fc.noarch requires serv
Re: Anaconda: good work!
From: "Bill Crawford" > That said, a blinking text for failure *would* make it stand out; in > particular the whole red/green thing only works well when it's one of those > standing out in contrast to the other. In cases where there's only a small > amount of coloured text, it's not all that helpful anyway. Tha's why I suggested that the background be green/red, since it provides a lot more color. Having other elements (like blinking) to differentiate the two would also be good, but I think they should be in addition, rather than replacing, the color background. -- Dimi Paun Lattica, Inc.
Re: Anaconda: good work!
On Thu, 2006-03-23 at 09:32 +0100, Hans Kristian Rosbach wrote: > On Wed, 2006-03-22 at 11:45 -0500, Jeremy Katz wrote: > > On Wed, 2006-03-22 at 10:12 -0500, Dimi Paun wrote: > > > First impression: anaconda really rocks! This is a > > > very solid, pleasant installer. > > > > Thanks -- always nice to see people not just flaming us :-P > > *snip* > > I have to agree that the new installer is very nice. > But despite this I really wish you will apply some bugfixes > to anaconda and release it in updates. This of course does > not help with FC5, but it will help those who use anaconda > to make new distros or respins such as I plan to do (either > official or unofficial). > > The complete lack of updates to anaconda in FC4 bugged me, > but I was too late in the cycle to make you reconsider. > This time around I hope you will atleast take obvious > bugfixes and maintain the FC5 anaconda atleast for a > little while. I'm not asking for new features, only fixes. > > -HK Are you asking this as a general idea or do you have specific bugs that you want to see fixed sooner? . Bugzilla links would be nice. Rahul
Re: Anaconda: good work!
On Thu, 2006-03-23 at 09:32 +0100, Hans Kristian Rosbach wrote: >> This time around I hope you will atleast take obvious >> bugfixes and maintain the FC5 anaconda atleast for a >> little while. I'm not asking for new features, only fixes. > >Are you asking this as a general idea or do you have specific bugs that >you want to see fixed sooner? . Bugzilla links would be nice. Yes, the general idea. I have no bugreports so far. Just thought I'd voice this before the maintainers just let go and completely focus on the one for FC6. I bet obvious fixes will be detected in the starting phase of developing for FC6 and those might be backported to FC5 aswell. -HK
Re: Anaconda: good work!
Hans Kristian Rosbach wrote: > On Wed, 2006-03-22 at 11:45 -0500, Jeremy Katz wrote: > > On Wed, 2006-03-22 at 10:12 -0500, Dimi Paun wrote: > > > First impression: anaconda really rocks! This is a > > > very solid, pleasant installer. > > > > Thanks -- always nice to see people not just flaming us :-P > > *snip* > > I have to agree that the new installer is very nice. > But despite this I really wish you will apply some bugfixes > to anaconda and release it in updates. This of course does > not help with FC5, but it will help those who use anaconda > to make new distros or respins such as I plan to do (either > official or unofficial). > > The complete lack of updates to anaconda in FC4 bugged me, > but I was too late in the cycle to make you reconsider. > This time around I hope you will atleast take obvious > bugfixes and maintain the FC5 anaconda atleast for a > little while. I'm not asking for new features, only fixes. If you're using anaconda in projects, it's probably worth following what happens in rawhide. New anaconda packages make it there all the time. -- David Cantrell Red Hat / Westford, MA
Re: Unable to mount cd & dvd after upgrading to FC5
On 3/23/06, Igor Jagec wrote: > > 1) if you tried to run it manually... does that mean it wasn't running > > already? > > Most likely it wasn't. hald not running..even in fc4.. is a problem that will result in a number of weird symptoms. Unfortunately "most likely" doesn't help narrow the underlying problem down, i was hoping for confirmation while you were experiencing the problem. > > > 2)is the dbus stuff running correctly? /sbin/service messagebus status > > [r...@munja ~]# /sbin/service messagebus status > dbus-daemon (pid 4147 1558) se izvršava... > > Which means it runs properly. BTW I tried to get english output with > 'export LANG=en_EN.ISO8859-1', and it didn't help for that command, and > for some it did (?). Never mind. > > > 3) does the outout of lshal show your hdc and hdd devices? > > [r...@munja ~]# lshal|grep hdc I wasn't asking for a grep.. because the grep won't extract all the information in the "block" of output that covers all the properties. I specifically ask for a "block" oufput and I tried to define what the beginning and end of the block looks like. Unfortunately you didn't seem to understand what I was asking for. > I didn't know how to provide you more information, but I hope that above > will help a bit. I saw on the redhat's news group that I'm not the only > one who experienced that problem. gnome-mount isnt crashing for me... and I don't see any bugreports about gnome-mount crashing so far reported in bugzilla by anyone. If this is a common occurance, and its a real bug, someone from those "news groups" needs to actually file a bug report if they want the crashing fixed. People can discuss crasher issues in forums and newsgroups and mailinglists forever.. but if its not filed in bugzilla..you can never be sure the developers who need to be aware of it will know about it. -jef
Re: Anaconda: good work!
On 3/23/06, Ralf Ertzinger wrote: > This step (what does that do, anyway?) is _slow_. 30 seconds per > schema on my 500MHz iBook. Updating gnome-games takes ages. I'd have to agree, I've seen the schema updates take quite a bit of time. So long in fact that I actually flip over and start a top to see if the rpm trasaction is stalled or not. -jef
Re: Anaconda: good work!
Paul Howarth wrote: > I upgraded a desktop box from FC4->FC5 yesterday, and my perception was > that the process was significantly slower (perhaps taking twice as > long?) as previous upgrades, e.g. FC3->FC4 on the same box. Can't > explain why that should be, unless of course it was to give me plenty of > time to read the release notes (which I was able to do in full during > the process) :-) I had the same experience. In a number of cases it took an uncomfortably long time (15 seconds?) for the installer to respond to a button click. -- Ian Pilcheri.pilc...@comcast.net
Re: Anaconda: good work!
> I have to agree that the new installer is very nice. > But despite this I really wish you will apply some bugfixes > to anaconda and release it in updates. This of course does > not help with FC5, but it will help those who use anaconda > to make new distros or respins such as I plan to do (either > official or unofficial). > > The complete lack of updates to anaconda in FC4 bugged me, > but I was too late in the cycle to make you reconsider. > This time around I hope you will atleast take obvious > bugfixes and maintain the FC5 anaconda atleast for a > little while. I'm not asking for new features, only fixes. We can't really do this easily. Making a new anaconda means making new first and second stage images, which means making new CD and DVD images, which means a huge headache for mirrors and for keeping track of all these versions of images floating around. One thing we can look into is the possibility of providing an updates.img from time to time that rolls up fixes, but this has its own set of problems. Doing something like that will make debugging in the future much harder because we'll have to ask people what updates they might be using in their installation when they report bugs. I can just see all sorts of problems resulting from that. The installer is special. - Chris
Re: Anaconda: good work!
On Wed, 2006-03-22 at 15:40 -0800, David Kewley wrote: > On Wednesday 22 March 2006 09:38, Dimi Paun wrote: > > From: "Jeremy Katz" > > > > * Disk change indicators > > > > > > This comes up from time to time, I'm just not fully convinced of how > > > useful it is. > > > > Coupled with a better time estimate, it would be useful, no doubt. > > Also copying the contents of the CDs to the HD before hand would be > > a very good idea too. But when that's not possible (due to free > > space constraints), the ticks would be nice :) > > A suggestion if this gets considered: dd the CD to a .iso file on disk, and > mount it loopback. That way you avoid slow CD seeks. The cost is that you then have to read every bit off the CD as opposed to just the ones for the packages you're installing as well as then needing to seek more on the disk when installing packages (since you're reading and writing from the same device). There's also a non-trivial disk space overhead Jeremy
Re: Anaconda: good work!
On Wed, 2006-03-22 at 21:29 +0100, Thomas Canniot wrote: > What I found sad in anaconda during the installation process is the > permanent showing of our brand new logo for about 20 min, while anaconda > is copying files. We had in previous version of FC, texts talking about > the distribution itself. These have disappeared, too bad really. The main thing needed here is having a group step up to create the content. fedora-marketing-love project maybe? :-) Jeremy
Core 5 and Samba issues
From a fresh install of FC5 I was joining my box to the office domain and came across a few issues. Let me start by saying, system-config-samba is a very good tool. It has some holes in functionality, but samba has always been a thorn in my side and this tool made it a whole lot easier. However, if I go to Server Settings and select "ADS" for authentication, when I enter my Realm it joins the realm line with the line that follows it in the smb.conf which causes problems even when the following line is a comment. Second, there seems to be an issue with SELinux policies and kerberos/samba. When I've properly configured /etc/krb5.conf and samba, then try to do a "net join", the action fails with the error "end point not connected". Even selecting "Disable SELinux" for smbd, nmbd, winbind and kerberos, the action fails. After "setenforce 0" the action succeeds. After the "net join" is complete, selinux can (apparently) be reset to enforcing. Finally, SWAT is evil. I used it to add a printer and it obliterated my config from system-config-samba. Is there somewhere to make a recommendation to the user to use one or the other, but not both? e.
Re: Anaconda: good work!
Bill Nottingham wrote: dragoran (drago...@feuerpokemon.de) said: I don't know if this is a feature or a bug but it seems like a bug: I tryed a harddisk install (the image was on /dev/md0 ) but I could'nt select it it only showed all /dev/sd(a|b)X but not /dev/md0 any reason for that or is this a bug? Well, depending on your point of view, I believe that's either a bug that's always existed, or a request for a feature enhancement. :) Bill no it seems that raid0 is broken in anaconda and/or fc5 kernel see: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=186312
Today' rawhide update
Is it just me or did we not get a rawhide update today? - David signature.asc Description: Dette er en digitalt underskrevet brevdel
Re: Anaconda: good work!
Garry Williams (gtwilli...@gmail.com) said: > > One thing that we really probably need to look at for FC6 > > is revisiting the time remaining algorithm. > > Yeah, it took about fifteen minutes after all packages were installed to > finish doing an update. That was long "1 minute remaining". :-) In prior releases, the upgrade procedure was: upgrade A remove old A upgrade B remove old C ... upgrade Z *** remove old Z With the switch to yum, it's now: upgrade A upgrade B upgrade C ... upgrade Z *** remove old A remove old B remove old C ... remove old Z The time algorithm and the progress bar were adopted for the first method - when you get to 'where all the packages are installed', you're at the point makred by '***' above. This obviously doesn't fit the new model very well. Bill
Re: Broken bind (Re: rawhide report: 20060321 changes)
Dnia 22-03-2006, śro o godzinie 13:42 -0800, Jesse Keating napisał(a):
[..]
> Tomasz, before (or if) you renamed the named.caching-nameserver.conf, did
> named prefer this file over your modified named.conf ?
No, named.caching-nameserver.conf was preffered but this doesn't metter
now because inti script from today bind-9.3.2-12.FC6 seems works
correctly.
Still IMO adding special handling for named.caching-nameserver.conf in
separated bind-config package is IMO wrong way (IMO best will be put
caching only configuration as default configura example .
BTW I found next bug in current bind. Iin libbind.pc in Libs: is
specyfied -L/usr/lib - this will break 64bit archs
Patch in attachment.
I second atachemnt is patch for init scrips (use tabs instead spaces,
removed trailing spaces and tabs, removed ";" on EOF; rolled in multipe
rm commands to single run with more than one file for remove).
In next attachment are bind.spec cleanups:
- removed trailig spaces and tabs,
- use tabs instead spaces,
- remove gcc and tar from BuildRequires (this is esential BR so specify
this directly isn't neccessary),
- simplified BuildRequires rules,
- s/textutils, fileutils/coreutils/ in Requires rules,
- removed not neccessatry specs conditions (%if in packages descriptions
and %post/%postun scripts can be ommited),
- simplifications in %build,
- s/\${RPM_BUILD_ROOT}/$RPM_BUILD_ROOT/
- %clean moved after %install and before all %post/%pre/%post/%postun
scripts,
- s,/usr/share,%{_datadir},
- cleanups in %prep,
- move unpack RFC docs to %prep (fixes for --short-circuit build),
- remove nit neccessary ";" in scripts and
- rewrited sdb %postun for use only sed (this allow minimize number of
Requires it will be also good rewrite %post script in sed because
using this allow remove from sdb Requires(post) mktemp and SELinux
utils because sed preserves SElinux context on files).
kloczek
Index: libbind.pc
===
RCS file: /cvs/dist/devel/bind/libbind.pc,v
retrieving revision 1.1
diff -u -r1.1 libbind.pc
--- libbind.pc 8 Mar 2006 18:21:32 - 1.1
+++ libbind.pc 23 Mar 2006 16:56:31 -
@@ -2,5 +2,5 @@
Name: libbind
Description:BIND 8 compatible DNS Resolver Library
Version: 9.3.2
-Libs: -L/usr/lib -lbind
+Libs: -lbind
Cflags: -I/usr/include/bind
\ No newline at end of file
Index: named.init
===
RCS file: /cvs/dist/devel/bind/named.init,v
retrieving revision 1.42
diff -u -r1.42 named.init
--- named.init 22 Mar 2006 17:55:16 - 1.42
+++ named.init 23 Mar 2006 16:59:34 -
@@ -1,12 +1,12 @@
#!/bin/bash
#
-# named This shell script takes care of starting and stopping
-# named (BIND DNS server).
+# named This shell script takes care of starting and stopping
+# named (BIND DNS server).
#
-# chkconfig: - 13 87
-# description: named (BIND) is a Domain Name Server (DNS) \
+# chkconfig: - 13 87
+# description: named (BIND) is a Domain Name Server (DNS) \
# that is used to resolve host names to IP addresses.
-# probe: true
+# probe: true
# Source function library.
. /etc/rc.d/init.d/functions
@@ -20,52 +20,53 @@
[ -r /etc/sysconfig/named ] && . /etc/sysconfig/named
if [ -n "$ROOTDIR" ]; then
- ROOTDIR=`echo $ROOTDIR | sed 's#//*#/#g;s#/$##'`;
- rdl=`/usr/bin/readlink $ROOTDIR`;
- if [ -n "$rdl" ]; then
- ROOTDIR="$rdl";
- fi;
+ ROOTDIR=`echo $ROOTDIR | sed 's#//*#/#g;s#/$##'`
+ rdl=`/usr/bin/readlink $ROOTDIR`
+ if [ -n "$rdl" ]; then
+ ROOTDIR="$rdl"
+ fi
fi
RETVAL=0
named='named'
if [[ "$ENABLE_SDB" = [yY1]* ]] && [ -x /usr/sbin/named_sdb ]; then
-named='named_sdb';
-fi;
+ named='named_sdb'
+fi
prog=$named
[ -x /usr/sbin/$named ] || exit 1
nmdcOption()
-{ let i=0;
- for a in $*;
- do ((++i));
- if [ $a = -c ]; then
- ((++i));
- eval 'echo $'$i;
- elif [[ $a = -c* ]]; then
- echo ${a#-c};
- fi;
- done;
+{
+ let i=0
+ for a in $*
+ do ((++i))
+ if [ $a = -c ]; then
+ ((++i))
+ eval 'echo $'$i
+ elif [[ $a = -c* ]]; then
+ echo ${a#-c}
+ fi
+ done
}
-named_c_option=`nmdcOption $OPTIONS`;
-named_conf=${named_c_option:-/etc/named.conf};
+named_c_option=`nmdcOption $OPTIONS`
+named_conf=${named_c_option:-/etc/named.conf}
if [ ! -r ${ROOTDIR}${named_conf} ] ; then
- if [ -z "$named_c_option" ] && [ -r ${ROOTDIR}/etc/named.caching-nameserver.conf ]; then
- named_conf='/etc/named.caching-nameserver.conf';
- OPTIONS="$OPTIONS -c ${named_conf}";
- else
- echo Locating $ROOTDIR/${named_conf} failed:
- failure
- echo;
- fi;
-fi;
+ if [ -z "$named_c_option" ] && [ -r ${ROOTDIR}/etc/named.caching-nameserver.conf ]; then
+ named_conf='/etc/named.caching-nameserver.conf'
+ OPTIONS="$OPTIONS -c ${named_conf}"
+ else
+ echo Locating $ROOTDIR/${named_conf} failed:
+ failure
+ echo
+ fi
+fi
start() {
-# Start daemons.
-echo -n
Re: Anaconda: good work!
On Thu, 2006-03-23 at 09:30 -0500, Jeff Spaleta wrote: > On 3/22/06, Thomas Canniot wrote: > > Maybe we could do something here that really helps people, newbies who > > are just installing their first linux distribution. I dreamt of an > > anaconda that helps newbies make their first steps in Fedora Core. > > Screw that... embedded game of nexuiz would be much better. > > As for the mock up... perhaps an embedded ogg video of desktop > interactions with localized closed caption text ala annodex. > > -jef"so what if anaconda would require a minimum of 2 gig of ram for > the video to play while anaconda does packaging actions."spaleta What about just slides like powerpoint (images)? Richard.
Re: gnome-screen-saver unlock artwork
On 3/23/06, Thomas J. Baker wrote: > Has anyone else seen this? that's awesome!! how to diagnose that lets see first guess would be some sort of gdm/gtk/gnome theming which could be related per user? Like the gnome splash theme.. isnt that per user configurable now.. and thus have some sort of systme default as well. anything in /var/log/message that seems to be associated with screensaver turning on or the lock dialog showing up? -jef
Removing dep for httpd from php?
Hi, I was wondering if removing the dep for httpd inside package php might be justified? Seen that in FC4 - and I guess it's in FC5 as well. When running a server without a webserver installed, you can perfectly install/use php anyhow. So why is that dep needed? Is there a chance to get it removed on the next release? :-) Regards, Stefan
Re: gnome-screen-saver unlock artwork
On Thu, 2006-03-23 at 09:50 -0500, Thomas J. Baker wrote: > I have some inconsistencies across several machines with the artwork for > the gnome screen saver unlock window that I thought was due to different > initial fc5 test installs. Now that I've done clean installs of fc5 > proper on the same systems the inconsistencies remain. On some systems I > have the themed fedora core 5 unlock window and on others I have a > simple grey box. I can't really do a screen shot because the screen is > locked. It doesn't appear to be user related because a new user on an > affected system still has just the grey box unlock dialog. The themed lock dialogs are per-screensaver. Only one screensaver currently provides a themed lock dialog. It was probably a bad idea to do it this way. It causes a lot of confusion. --Ray
Re: gnome-screen-saver unlock artwork
On Thu, 2006-03-23 at 14:19 -0500, Ray Strode wrote: > On Thu, 2006-03-23 at 09:50 -0500, Thomas J. Baker wrote: > > I have some inconsistencies across several machines with the artwork for > > the gnome screen saver unlock window that I thought was due to different > > initial fc5 test installs. Now that I've done clean installs of fc5 > > proper on the same systems the inconsistencies remain. On some systems I > > have the themed fedora core 5 unlock window and on others I have a > > simple grey box. I can't really do a screen shot because the screen is > > locked. It doesn't appear to be user related because a new user on an > > affected system still has just the grey box unlock dialog. > The themed lock dialogs are per-screensaver. Only one screensaver > currently provides a themed lock dialog. It was probably a bad idea to > do it this way. It causes a lot of confusion. > > --Ray > Thanks for the explanation. The Floating Fedora Bubbles is the one with the themed unlock dialog. At least there's a logical explanation and no gremlins! tjb -- === | Thomas Baker email: t...@unh.edu| | Systems Programmer | | Research Computing Center voice: (603) 862-4490 | | University of New Hampshire fax: (603) 862-1761 | | 332 Morse Hall | | Durham, NH 03824 USA http://wintermute.sr.unh.edu/~tjb | ===
Re: rawhide report: 20060323 changes
On Thu, 2006-03-23 at 09:56 -0500, Build System wrote: > > > Updated Packages: > > (none) There were some issues with today's rawhide tree spin. A correct rawhide report email will be coming out in a minute. -- Dennis
rawhide report: 20060323 changes
Updated Packages: aspell-12:0.60.3-6 -- * Wed Mar 22 2006 Ivana Varekova - 12:0.60.3-6 - remove .la files (bug 184184) * Thu Mar 02 2006 Ivana Varekova - 12:0.60.3-5 - update aspell man page (bug 183205) * Tue Feb 21 2006 Ivana Varekova - 12:0.60.3-4 - fix multilib file conflict bash-3.1-10 --- * Thu Mar 23 2006 Tim Waugh 3.1-10 - Patchlevel 14. * Thu Mar 02 2006 Tim Waugh 3.1-9 - Fixed duplicate documentation of ulimit '-x' option introduced by ulimit patch (bug #183596). * Tue Feb 21 2006 Tim Waugh 3.1-8 - Patchlevel 10. beagle-0.2.3-5 -- * Tue Mar 21 2006 Alexander Larsson 0.2.3-5 - Rebuild fc5 update in rawhide bind-30:9.3.2-12.FC6 * Wed Mar 22 2006 Jason Vas Dias - 30:9.3.2-12 - fix typo in initscript - fix Requires(post): policycoreutils in sub-packages cpio-2.6-14.FC5 --- * Wed Mar 22 2006 Peter Vrabec 2.6-14.FC5 - FC5 update * Wed Mar 15 2006 Peter Vrabec 2.6-13 - merge toAsciiError.patch with writeOutHeaderBufferOverflow.patch - merge largeFileGrew.patch with lfs.patch - fix large file support, cpio is able to store files<8GB in 'old ascii' format (-H odc option) - adjust warnings.patch * Tue Mar 14 2006 Peter Vrabec 2.6-12 - fix warn_if_file_changed() and set exit code to #1 when cpio fails to store file > 4GB (#183224) cups-1:1.2-0.1.b2.3 --- * Thu Mar 23 2006 Tim Waugh 1:1.2-0.1.b2.3 - Update to svn snapshot. No longer need users or policy patches. expect-5.43.0-4 --- * Fri Feb 24 2006 David Cantrell - 5.43.0-4 - Patch expLogChannelOpen() to create files with 0666 permissions (#182724) gnome-icon-theme-2.14.2-2 - * Wed Mar 22 2006 Matthias Clasen 2.14.2-2 - Update to 2.14.2 - Add symlinks to make application/xml work kernel-2.6.16-1.2083_FC6 * Wed Mar 22 2006 Dave Jones - 2.6.16-git5 * Wed Mar 22 2006 David Woodhouse - Update the bcm43xx driver to make it work nicely with initscripts and NetworkManager without user intervention. * Tue Mar 21 2006 Dave Jones - 2.6.16-git3 - Improve spinlock scalability on big machines. m17n-db-1.3.3-2 --- * Thu Mar 09 2006 Jens Petersen - 1.3.3-2 - Bengali input maps fixes (runab) - map Probhat '*' key to an alternate sequence since glyph missing (#179821) - more itrans cleanup (#182227) - add icon for Tamil99 (aalam) man-pages-fr-0.9.7-14 - * Thu Mar 23 2006 Karsten Hopp 0.9.7-14 - remove pages that conflict with vim man-pages-it-0.3.0-17 - * Thu Mar 23 2006 Karsten Hopp 0.3.0-17 - remove vim.1, provided by the vim-common package openoffice.org-1:2.0.2-5.4.3 * Mon Mar 13 2006 Caolan McNamara - 1:2.0.2-5.4 - ooo#59997# replacement opens___.ttf updated - drop integrated openoffice.org-2.0.0.ooo56651.sw.rtfcrash.patch - drop integrated openoffice.org-1.9.114.ooo51718.rpath.patch - add openoffice.org-2.0.2.ooo63155.sfx2.badscript.patch for rh#185390# - rh#181900# rename Bengali langpack - drop pagein swappiness foo - drop nearly 9 megs of afms and ppds perl-DBD-Pg-1.47-1 -- * Wed Mar 22 2006 Jason Vas Dias - 1.47-1 - Upgrade to upstream version 1.47 perl-HTML-Parser-3.51-1.FC6 --- * Wed Mar 22 2006 Jason Vas Dias - 3.51-1 - upgrade to 3.51 perl-Net-DNS-0.57-1 --- * Wed Mar 08 2006 Jason Vas Dias - 0.57-1 - Upgrade to upstream version 0.57 php-pear-1:1.4.6-2.1 * Wed Mar 22 2006 Joe Orton 1:1.4.6-2.1 - update to XML_RPC 1.4.5 (#186140) postgresql-odbc-08.01.0200-2 * Wed Mar 22 2006 Tom Lane 08.01.0200-2 - Change library name back to psqlodbc.so, because it appears that upstream will revert to that name in next release; no point in thrashing the name. - Include documentation files unaccountably omitted before (bug #184158) qt-1:3.3.6-1 * Mon Mar 20 2006 Than Ngo 1:3.3.6-1 - update to 3.3.6 - adapt qt-x11-immodule-unified-qt3.3.5-20060318 to qt-3.3.6 - remove set of fixes for the immodule patch, included in qt-x11-immodule-unified-qt3.3.5-20060318 - remove 0051-qtoolbar_77047.patch, qt-x11-free-3.3.4-assistant_de.patch, qt-x11-free-3.3.5-warning.patch, included in new upstream selinux-policy-2.2.25-2 --- * Wed Mar 22 2006 Dan Walsh 2.2.25-2 - Fix pam_console handling of usb_device - dontaudit logwatch reading /mnt dir sendmail-8.13.6-1 - * Wed Mar 22 2006 Thomas Woerner 8.13.6-1 - new version 8.13.6 (fixes VU#834865) - dropped libmilter-sigwait patch (fixed in 8.13.6) shadow-utils-2:4.0.14-5.FC5 --- * Wed Mar 22 2006 Peter Vrabec 2:4.0.14-5.FC5 * FC5 update * Fri Mar 10 2006 Peter Vrabec 2:4.0.14-4 - fix lrename() function to handle relative symlinks too * Tue Mar 07 2006 Peter Vrabec 2:4.0.14-3 - set default umask to 077 in login.defs shared-mime-info-0.17-2 ---
Kernel for SMP VIA C3 machines
I have a couple of VT310-DP boards (together in a little 1U case, nice) and found that FC-5 won't do SMP on them. (The i686 kernels won't run on this chip.) What's the cleanest way to add an i586-smp (or better yet, MVIAC3_2-smp) build target to the existing kernel SRPM? Currently I'm working from Fedora CVS (FC-5 branch) which looks like it has reorganized the config generation a bit. - J<
Re: Anaconda: good work!
On 3/23/06, Hans Kristian Rosbach wrote: > Just thought I'd voice this before the maintainers just let go > and completely focus on the one for FC6. I bet obvious fixes > will be detected in the starting phase of developing for FC6 > and those might be backported to FC5 aswell. Installer bugs happen with pretty much each release. fixed boot.isos are usually made available as links in bugzilla tickets as issues are addressed and I believe there is a mechanism which is applied to incorporate fixes into the mirrors so people doing network installs can avoid some problems. But at no point have I ever seen any discussion at any time which suggests the release team is interested in spinning up replacement isos and distributing them. And quite frankly I don't think this is the most appropriate time to suggest a change in the release model used. This is the sort of thing that should be debated during the testing phase, so plans can be in place to support respins if you are able to convince the people who have to do them that its a good idea. I don't think your request for fc5 anaconda updates post release day is going to change any minds as to the support tradeoffs associated with official respins that incorporate installer changes. -jef"too little too late"spaleta
boot images gone?
What happened to images/* and isolinux/ on rawhide? They seem to have vanished last night. Or is it just my mirror. ../C
Re: Anaconda: good work!
> "CL" == Chris Lumens writes: CL> Making a new anaconda means making new first and second stage CL> images, which means making new CD and DVD images, which means a CL> huge headache for mirrors and for keeping track of all these CL> versions of images floating around. The thing is, if you're spinning your own distro then you should be able to figure out how to rebuild the anaconda RPM out of current CVS. The last time I had to do it it took only a few additional minutes to build an anaconda with fixes, along with other packages (kudzu in this case) that needed updates. The problems crop up when the current development version becomes unsuitable for installing a current tree (as happened with the switch to yum). At that point it might be nice if the anaconda developers would consider checking any critical fixes into the release branch for those of us who regularly respin our distros, but I don't see the point in actually doing releases. - J<
