Description of problem: The FC6 version of Anaconda seems to pick the i586 kernel instead of the i686 one it's supposed to on some P4 and Centrino systems. I had this happen to me with an install of Rawhide from the 21st (which is essentially FC6), Guillaume Kulakowski had this happen (see URL with description in French) on a Centrino with the official FC6 release. Version-Release number of selected component (if applicable): anaconda-11.1.1.3-1 How reproducible: No idea. Steps to Reproduce: 1. Install FC6 on an affected system and check the kernel architecture. Actual results: The kernel is an i586 kernel. Expected results: The kernel is an i686 kernel. Additional info: My (brutal) fix: 1. wget the i686 kernel 2. rpm -e kernel 3. rpm -ivh kernel-*.i686.rpm I don't think rpm -e on the current kernel is that great an idea though. I risked it because I couldn't think of a better solution which doesn't screw up the RPM database and/or the file system in one way or the other.
(Note that this is a new computer with a fresh install, not the one where the Smart PM decided to pick an i586 kmod for some reason. That one was part of #192775.)
Please attach the output of cat /proc/cpuinfo and also if you do mv /etc/rpm/platform /etc/rpm/platform.bak; rpm --showrc | head; mv \ /etc/rpm/platform.bak /etc/rpm/platform
Same here on a Dell Latitude D800 (which has a Pentium M 1.6, see attached cpuinfo): root@gibraltar:~> mv /etc/rpm/platform /etc/rpm/platform.bak; rpm --showrc | head; mv \ > /etc/rpm/platform.bak /etc/rpm/platform ARCHITECTURE AND OS: build arch : i386 compatible build archs: pentium3 i686 i586 i486 i386 noarch fat build os : Linux compatible build os's : Linux install arch : pentium3 install os : Linux compatible archs : pentium3 i686 i586 i486 i386 noarch fat compatible os's : Linux
Created attachment 139222 [details] /proc/cpuinfo of Dell Latitude D800
Same thing happened on Toshiba Satellite L25, which has Celeron M. # mv /etc/rpm/platform /etc/rpm/platform.bak; rpm --showrc | head; mv \ > /etc/rpm/platform.bak /etc/rpm/platform ARCHITECTURE AND OS: build arch : i386 compatible build archs: i686 i586 i486 i386 noarch fat build os : Linux compatible build os's : Linux install arch : i686 install os : Linux compatible archs : i686 i586 i486 i386 noarch fat compatible os's : Linux
Created attachment 139239 [details] cpuinfo for Toshiba Satellite L25
*** Bug 211790 has been marked as a duplicate of this bug. ***
Please look at Bug 211790. I had the same problem on Toshiba Satellite 1410-614 [lukasz@evil ~]$ cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 15 model : 2 model name : Mobile Intel(R) Celeron(R) CPU 1.80GHz stepping : 7 cpu MHz : 1794.275 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe up cid bogomips : 3592.16 [root@evil ~]# uname -m i686
Before I try "rpm -e kernel" as above, is there a safer way to replace the 586 with the 686 version? Or is there going to be a kernel update soon (in which case I could just boot into the new one first and then erase and replace the old one)?
Re: #c9 rpm -e --nodeps kernel and then rpm -ivh kernel.i686 will work to replace it make sure you back up /boot/grub/grub.conf first as you may have to rebuild it manually afterwards. grubby gets weird when your system gets down to zero kernels ;-)
I did run into the grubby bug too. Can someone paste the default /boot/grub/grub.conf file here? Thanks.
For me, it said it won't do anything because there wouldn't be any kernel left, and then after installing the new one, the grub.conf was valid.
Any reason why rpm -Fvh kernel-2.6.18-1.2798.fc6.i686.rpm wouldn't work as well? This bug isn't a big deal to me - the machine seems fast enough, so I'll probably wait for a kernel update. Another possibility might be to install one of the earlier development kernels, boot into that, then replace the release kernel.
Never mind - I tried the above command on kernel-devel first, and it doesn't consider the 686 version to be newer than the 586 one. However, rpm -Uvh --replacepkgs kernel-devel-2.6.18-1.2798.fc6.i686.rpm works. The same command on the kernel gives rpm -Uvh --replacepkgs kernel-2.6.18-1.2798.fc6.i686.rpm Preparing... ########################################### [100%] file /boot/System.map-2.6.18-1.2798.fc6 from install of kernel-2.6.18-1.2798.fc6 conflicts with file from package kernel-2.6.18-1.2798.fc6 file /boot/config-2.6.18-1.2798.fc6 from install of kernel-2.6.18-1.2798.fc6 conflicts with file from package kernel-2.6.18-1.2798.fc6 file /boot/symvers-2.6.18-1.2798.fc6.gz from install of kernel-2.6.18-1.2798.fc6 conflicts with file from package kernel-2.6.18-1.2798.fc6 file /boot/vmlinuz-2.6.18-1.2798.fc6 from install of kernel-2.6.18-1.2798.fc6 conflicts with file from package kernel-2.6.18-1.2798.fc6 The next logical thing to try would be rpm -Uvh --replacefiles --replacepkgs kernel-2.6.18-1.2798.fc6.i686.rpm but I don't have the nerve to try it right now.
I run into the same problem on P4 3.0E (Prescott, Socket 478)
Created attachment 139341 [details] cpuinfo for my (Kevin Kofler's) P4 Here's my cpuinfo (attached) and the output from Paul Nasrat's line: build arch : i386 compatible build archs: pentium4 pentium3 i686 i586 i486 i386 noarch fat build os : Linux compatible build os's : Linux install arch : pentium4 install os : Linux compatible archs : pentium4 pentium3 i686 i586 i486 i386 noarch fat compatible os's : Linux (The cpuinfo is without HT, but I had HT enabled when I did the install, just in case this matters. So the cpuinfo during install had "siblings: 2" where it now has "siblings: 1", I assume.)
FC6 VMware guest (single CPU enabled) on FC5 Host processor : 0 vendor_id : GenuineIntel cpu family : 15 model : 6 model name : Intel(R) Pentium(R) D CPU 3.40GHz stepping : 8 cpu MHz : 3399.516 cache size : 2048 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 6 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss lm constant_tsc up pni ds_cpl cx16 lahf_lm bogomips : 6828.41 $ uname -m => i686 kernel.i586 installed
I note that this resembles the problem described in bug 186378, which has a number of other bugs marked as duplicates of it. The only way I was ever able to get an i686 kernel installed under FC5 was to make a custom repository with no i586 kernels in it. I tried it on many machines of differing hardware. But FC6test2 installed fine for me, so I figured the bug was fixed. I have not yet tried an FC6 install.
its not just intels that this is happening with. [root@localhost yum.repos.d]# cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 6 model : 6 model name : AMD Athlon(tm) XP 2100+ stepping : 2 cpu MHz : 1726.068 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 mtrr pge mca cmov pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow up ts bogomips : 3453.76 [root@localhost yum.repos.d]# uname -m i686 [root@localhost yum.repos.d]# yum -d0 list kernel Installed Packages kernel.i586 2.6.18-1.2798.fc6 installed Available Packages kernel.i686 2.6.18-1.2798.fc6 core
Happens on my Pentium 4 1.8GHz Dell Inspiron 8200 cpuinfo -------- processor : 0 vendor_id : GenuineIntel cpu family : 15 model : 2 model name : Intel(R) Pentium(R) 4 Mobile CPU 1.80GHz stepping : 4 cpu MHz : 1200.000 cache size : 512 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm up bogomips : 2395.50 rpm --showrc ---------------- ARCHITECTURE AND OS: build arch : i386 compatible build archs: pentium4 pentium3 i686 i586 i486 i386 noarch fat build os : Linux compatible build os's : Linux install arch : pentium4 install os : Linux compatible archs : pentium4 pentium3 i686 i586 i486 i386 noarch fat compatible os's : Linux Kevin Kofler, would you please elaborate on the steps you used to upgrade to the i686 kernel for those of us not familiar with wget and rpm? I would appreciate a listing of the full syntax please. Been reading doc and searching google but I'm still not getting it right. Thanks.
So far, all the affected systems we got cpuinfos of are CPU family 6 or 15.
This seems to work: %rpm -Uvh --force ~bkoz/incoming/kernel-2.6.18-1.2798.fc6.i686.rpm Preparing... ########################################### [100%] 1:kernel ########################################### [100%] For the record, this was on an upgrade for: %cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 13 model name : Intel(R) Pentium(R) M processor 1.86GHz stepping : 8 cpu MHz : 1862.289 cache size : 2048 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss tm pbe nx up est tm2 bogomips : 3728.26
Created attachment 139375 [details] /proc/cpuinfo and platform output
I also ran into this problem on an AMD64 X2 3800+. It was causing strangeness when attempting to install livna's nvidia drivers (arch i586 vs i686). My fix was to swap out the i586 kernel for the i686 with yum, rather than rpm directly, and it worked just fine: 1) yum remove kernel kernel-devel 2) yum install kernel.i686 kernel-devel.i686 3) yum install (the list of a dozen odd i386 kernel depencies removed in step #1, such as "gnome-volume-manager" and "compiz")
Again, 6 and 15. Is this because they use no other values for CPU family or is there a correlation between CPU family and this bug?
I got the same on a Core Duo laptop, here's the cpu information: processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 14 model name : Genuine Intel(R) CPU T2400 @ 1.83GHz stepping : 8 cpu MHz : 1000.000 cache size : 2048 KB physical id : 0 siblings : 2 core id : 0 cpu cores : 2 fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 10 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc pni monitor vmx est tm2 xtpr bogomips : 3665.68 processor : 1 vendor_id : GenuineIntel cpu family : 6 model : 14 model name : Genuine Intel(R) CPU T2400 @ 1.83GHz stepping : 8 cpu MHz : 1000.000 cache size : 2048 KB physical id : 0 siblings : 2 core id : 1 cpu cores : 2 fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 10 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc pni monitor vmx est tm2 xtpr bogomips : 3661.87
Confirmed that Jason's workaround in comment #24 above works well. Rebooted with no problems, now using i686 kernel. Have been able to load i686 nVidia kmods from livna with no problems. Thanks Jason!
Created attachment 139421 [details] cpuinfo
I successfully used rpm -Uvh --replacefiles --replacepkgs kernel-2.6.18-1.2798.fc6.i686.rpm to replace the kernel. The rpm man page says that --force is equivalent to all of --replacepkgs, --replacefiles, and --oldpackage, and since this isn't an older package, it should be equivalent to --force. My grub.conf for a dual-boot box was bytewise identical before and after, though the modification time was updated. BTW, my CPU family is 15. Is there anyone with CPU family 6 or 15 who did NOT have this happen? And are the vast majority of CPUs in one of these two families?
I just noticed that my home system has an i586 kernel as well: root@wombat:~> cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 6 model : 4 model name : AMD Athlon(tm) processor stepping : 4 cpu MHz : 1400.132 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 mtrr pge mca cmov pat pse36 mmx fxsr syscall mmxext 3dnowext 3dnow up bogomips : 2803.04 root@wombat:~> mv /etc/rpm/platform /etc/rpm/platform.bak; rpm --showrc | head; mv \ > /etc/rpm/platform.bak /etc/rpm/platform ARCHITECTURE AND OS: build arch : i386 compatible build archs: athlon i686 i586 i486 i386 noarch fat build os : Linux compatible build os's : Linux install arch : athlon install os : Linux compatible archs : athlon i686 i586 i486 i386 noarch fat compatible os's : Linux
Same thing on a vmware-workstation guest running on a Linux host. The host is a Dell Optiplex GX300. FC6 is installed via "kickstart" using a ks.cfg file nearly identical to the one I used with FC5 (and which correctly installed the i686 kernel) The /proc/cpuinfo seen by the guest is: processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 8 model name : Pentium III (Coppermine) stepping : 8 cpu MHz : 794.674 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 mmx fxsr sse up bogomips : 1606.53
Ditto, ditto, and ditto. Only mine happens on an Apple Macbook Pro (Intel Core Duo processor). processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 14 model name : Genuine Intel(R) CPU T2600 @ 2.16GHz stepping : 8 cpu MHz : 2306.463 cache size : 2048 KB physical id : 0 siblings : 2 core id : 0 cpu cores : 2 fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 10 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc pni monitor vmx est tm2 xtpr bogomips : 4333.25 processor : 1 vendor_id : GenuineIntel cpu family : 6 model : 14 model name : Genuine Intel(R) CPU T2600 @ 2.16GHz stepping : 8 cpu MHz : 2306.463 cache size : 2048 KB physical id : 0 siblings : 2 core id : 1 cpu cores : 2 fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 10 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc pni monitor vmx est tm2 xtpr bogomips : 4330.92
Can someone please add this to http://fedoraproject.org/wiki/Bugs/FC6Common ? Judging by the number of comments, it seems to be a pretty common issue.
Confirmed on a Dell Latitude X1 processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 13 model name : Intel(R) Pentium(R) M processor 1.10GHz stepping : 6 cpu MHz : 1097.377 cache size : 2048 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr mce cx8 apic mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss tm pbe up est tm2 bogomips : 2197.37
WARNING: Back up your /boot/grub/grub.conf before removing your only kernel, especially if you have a non-Fedora OS listed there too. Normally, grubby refuses to remove the only kernel entry, but if you have other OSes, it will remove it, and then it won't find a template to base the new entry on when you reinstall the kernel.
When I reinstalled using --replacefiles --replacepkgs (equivalent to --force here), as opposed to removing and then installing, it preserved the changes I had made (removing rhgb and quiet), and also the fact that it was set to dual-boot with XP. So this route appears safer than the 2-step process, though one should certainly back up grub.conf anyway, and then immediately check it afterwards, before rebooting.
*** Bug 212406 has been marked as a duplicate of this bug. ***
Oddly, the i686 kernel-headers get installed. Hmmm. [root@bilbo ~]# rpm -q kernel-devel kernel kernel-devel-2.6.18-1.2798.fc6.i686 kernel-2.6.18-1.2798.fc6.i586 [root@bilbo RPMS]# cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 15 model : 28 model name : AMD Sempron(tm) Processor 3100+ stepping : 0 cpu MHz : 1809.265 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt 3dnowext 3dnow up lahf_lm ts fid vid ttp bogomips : 3620.87
I did not post that before, but here is my cat /proc/cpuinfo output (I use HT, so there are 2 cpus): processor : 0 vendor_id : GenuineIntel cpu family : 15 model : 3 model name : Intel(R) Pentium(R) 4 CPU 3.00GHz stepping : 4 cpu MHz : 3014.700 cache size : 1024 KB physical id : 0 siblings : 2 core id : 0 cpu cores : 1 fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 5 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe constant_tsc pni monitor ds_cpl cid xtpr bogomips : 6033.62 processor : 1 vendor_id : GenuineIntel cpu family : 15 model : 3 model name : Intel(R) Pentium(R) 4 CPU 3.00GHz stepping : 4 cpu MHz : 3014.700 cache size : 1024 KB physical id : 0 siblings : 2 core id : 0 cpu cores : 1 fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 5 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe constant_tsc pni monitor ds_cpl cid xtpr bogomips : 6029.49 Also, I worked around the bug as suggested in comment #29.
The steps in comment #24 also worked for me (thanks Jason Farrell). Here they are again: 1) yum remove kernel kernel-devel 2) yum install kernel.i686 kernel-devel.i686 3) yum install compiz gnome-volume-manager systemtap gnome-session pcmciautils (NOTE: your step 3 might be different, see post #24 for more info) Here's my cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 6 model : 10 model name : AMD Athlon(tm) XP 2600+ stepping : 0 cpu MHz : 1921.029 cache size : 512 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow up ts bogomips : 3845.11
(In reply to comment #33) > Can someone please add this to http://fedoraproject.org/wiki/Bugs/FC6Common ? > Judging by the number of comments, it seems to be a pretty common issue. I have done so.
*** Bug 212091 has been marked as a duplicate of this bug. ***
Strangely, FC6 installed the 586 kernel package but the 686 kernel-PAE package on my machine.
There is no i586 or i386 kernel-PAE, so it can't pick the wrong one there. What is interesting is that it correctly picks the i686 glibc and openssl (at least it did for me), but the wrong kernel.
I did run in the same problem, this resulted in an error when the cpufreq module was loaded, so frequency scaling was broken. [~]$ cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 9 model name : Intel(R) Pentium(R) M processor 1300MHz stepping : 5 cpu MHz : 600.000 cache size : 1024 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr mce cx8 mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 tm pbe up est tm2 bogomips : 1199.90 This is a Thinkpad R40 2722-3GG
Installed Packages kernel.i686 2.6.18-1.2798.fc6 installed kernel-devel.i686 2.6.18-1.2798.fc6 installed my kernel is all-right but acpi-cpufreq still fails to load. I notice my /etc/cpuspeed is missing driver I tried adding speedstep-centrino driver line, but also need to get this driver in /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver, else getting [root@gesho etc]# service cpuspeed restart cat: /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver: No such file or directory FATAL: Module speedstep_centrino not found. /etc/cpuspeed.conf VMAJOR=1 VMINOR=1 # uncomment this and set to the name of your CPUFreq module #DRIVER="powernow-k7" # Let background (nice) processes speed up the cpu OPTS="$OPTS -n" # Add your favorite options here #OPTS="$OPTS -s 0 -i 10 -r" # uncomment and modify this to check the state of the AC adapter #OPTS="$OPTS -a /proc/acpi/ac_adapter/*/state" # uncomment and modify this to check the system temperature #OPTS="$OPTS -t /proc/acpi/thermal_zone/*/temperature 75"
dunno if this helps http://www.mail-archive.com/linux-acpi@vger.kernel.org/msg03458.html
Here is another, older machine example of the same issue: $ cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 8 model name : Pentium III (Coppermine) stepping : 6 cpu MHz : 996.811 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 mtrr pge mca cmov pat pse36 mmx fxsr sse up bogomips : 1995.21
Same here on a 5 years old Toshiba Satellite 3000-100 notebook: [root@notebook ~]# cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 8 model name : Pentium III (Coppermine) stepping : 10 cpu MHz : 847.452 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 mtrr pge mca cmov pat pse36 mmx fxsr sse up bogomips : 1696.40 [root@notebook ~]# mv /etc/rpm/platform /etc/rpm/platform.bak; rpm --showrc | head; mv /etc/rpm/platform.bak /etc/rpm/platform ARCHITECTURE AND OS: build arch : i386 compatible build archs: pentium3 i686 i586 i486 i386 noarch fat build os : Linux compatible build os's : Linux install arch : pentium3 install os : Linux compatible archs : pentium3 i686 i586 i486 i386 noarch fat compatible os's : Linux
# rpm -qa --queryformat "%{NAME}-%{VERSION}-%{ARCH}\n" | grep kernel kernel-devel-2.6.18-i586 kernel-headers-2.6.18-i386 kernel-2.6.18-i586 # cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 6 model : 10 model name : AMD Athlon(tm) XP 2000+ stepping : 0 cpu MHz : 1533.489 cache size : 512 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow up ts bogomips : 3068.72
What worked for me: rpm -ev kernel-devel #just to keep things simple # backup /boot/grub/grub.conf # download the correct kernel and kernel-devel rpm -Uvh --force kernel-2.6.18-1.2798.fc6.i686.rpm # restore grub.conf if necessary - wasn't for me reboot rpm -ivh kernel-devel-2.6.18-1.2798.fc6.i686.rpm
[sglaser@compaqlaptop ~]$ cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 15 model : 4 model name : AMD Athlon(tm) 64 Processor 3000+ stepping : 8 cpu MHz : 800.000 cache size : 1024 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext lm 3dnowext 3dnow up ts fid vid ttp bogomips : 1597.26 ARCHITECTURE AND OS: build arch : i386 compatible build archs: athlon i686 i586 i486 i386 noarch fat build os : Linux compatible build os's : Linux install arch : athlon install os : Linux compatible archs : athlon i686 i586 i486 i386 noarch fat compatible os's : Linux
#29 tip works good. processor : 0 vendor_id : AuthenticAMD cpu family : 15 model : 4 model name : AMD Athlon(tm) 64 Processor 3700+ stepping : 10 cpu MHz : 800.000 cache size : 1024 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext lm 3dnowext 3dnow up ts fid vid ttp bogomips : 1597.26
Wrong kernel installed here, family 15. processor : 0 vendor_id : GenuineIntel cpu family : 15 model : 2 model name : Intel(R) Pentium(R) 4 CPU 3.20GHz stepping : 9 cpu MHz : 3207.394 cache size : 512 KB physical id : 0 siblings : 2 core id : 0 cpu cores : 1 fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe cid xtpr bogomips : 6419.36
Same problem on ASUS A6K-Q034 [lystor@lystor SPECS]$ cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 15 model : 44 model name : Mobile AMD Sempron(tm) Processor 3100+ stepping : 2 cpu MHz : 800.000 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt 3dnowext 3dnow up pni lahf_lm ts fid vid ttp tm stc bogomips : 1600.97
AMD Duron 1300 CPU and the same problem here. The following taken after rebooted with kernel.i686: ARCHITECTURE AND OS: build arch : i386 compatible build archs: athlon i686 i586 i486 i386 noarch fat build os : Linux compatible build os's : Linux install arch : athlon install os : Linux compatible archs : athlon i686 i586 i486 i386 noarch fat compatible os's : Linux # cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 6 model : 7 model name : AMD Duron(tm) stepping : 1 cpu MHz : 1295.673 cache size : 64 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 mtrr pge mca cmov pat pse36 mmx fxsr sse syscall mp mmxext 3dnowext 3dnow up ts bogomips : 2593.84
*** Bug 213413 has been marked as a duplicate of this bug. ***
Thinkpad T42, got i586 not i686. ARCHITECTURE AND OS: build arch : i386 compatible build archs: i686 i586 i486 i386 noarch fat build os : Linux compatible build os's : Linux install arch : i686 install os : Linux compatible archs : i686 i586 i486 i386 noarch fat compatible os's : Linux processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 13 model name : Intel(R) Pentium(R) M processor 1.80GHz stepping : 6 cpu MHz : 599.504 cache size : 2048 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr mce cx8 mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss tm pbe up est tm2 bogomips : 1200.24
Created attachment 140294 [details] cpuinfo, rpm info, and dmidecode output The attached output is taken after upgrading to the i686 kernel. BTW - would it be more informative if this info came from booting the first CD in rescue mode and grabbing some output? Then the actual anaconda run-time kernel and software envirnoment would be active.
*** Bug 211768 has been marked as a duplicate of this bug. ***
i686 installed instead of i586. /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 15 model : 1 model name : Intel(R) Pentium(R) 4 CPU 1.60GHz stepping : 2 cpu MHz : 1616.030 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm up bogomips : 3235.40 ARCHITECTURE AND OS: build arch : i386 compatible build archs: pentium4 pentium3 i686 i586 i486 i386 noarch fat build os : Linux compatible build os's : Linux install arch : pentium4 install os : Linux compatible archs : pentium4 pentium3 i686 i586 i486 i386 noarch fat compatible os's : Linux
One more for the pot! Processor details below. I have got it sorted following the suggestion in comment #51 (thanks!), tho there were some loose ends, notably that on rebooting, the X server wouldn't start. Followed the instructions on screen to reconfigure. But it might be a good idea to disable the nvidia driver first, before rebooting. Also needed to reset the default language and keyboard. [root@dads james]# cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 15 model : 4 model name : Intel(R) Celeron(R) CPU 2.66GHz stepping : 9 cpu MHz : 2660.986 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 5 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe lm constant_tsc up pni monitor ds_cpl tm2 cid cx16 xtpr bogomips : 5326.02
I found a solution, you need to start Fedora Core 6 install with this command: linux i686 this instructs installer to use i686 kernel.
The same problem on a HP compaq nx7010. The installed kernel was the correct one (i686), but the kernel-devel package was the one for i586. cpuinfo: [root@s71-samhein ~]# cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 13 model name : Intel(R) Pentium(R) M processor 1.70GHz stepping : 6 cpu MHz : 1694.562 cache size : 2048 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr mce cx8 mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss tm pbe up est tm2 bogomips : 3391.38
Idem on HP Compaq nx7400 [root@localhost Desktop]# cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 14 model name : Genuine Intel(R) CPU T2300 @ 1.66GHz stepping : 8 cpu MHz : 1662.721 cache size : 2048 KB physical id : 0 siblings : 2 core id : 0 cpu cores : 2 fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 10 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc pni monitor est tm2 xtpr bogomips : 3327.52 processor : 1 vendor_id : GenuineIntel cpu family : 6 model : 14 model name : Genuine Intel(R) CPU T2300 @ 1.66GHz stepping : 8 cpu MHz : 1662.721 cache size : 2048 KB physical id : 0 siblings : 2 core id : 1 cpu cores : 2 fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 10 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc pni monitor est tm2 xtpr bogomips : 3324.57
I am maintaining Linux on several computers within a school. To install FC6 I've used a kickstart script so that FC6 can be installed without user-intervention. I'm also experiencing this bug on some computers. The strange thing is that all computers are the same (they're all HP Compaq's with a P4 2.8Ghz proc; don't know the product number out of my head) and this issue shows up randomly on some computers. After restarting the kickstart installation, the problem sometimes gets solved, sometimes it needs another re-installation.
ARCHITECTURE AND OS: build arch : i386 compatible build archs: pentium3 i686 i586 i486 i386 noarch fat build os : Linux compatible build os's : Linux install arch : pentium3 install os : Linux compatible archs : pentium3 i686 i586 i486 i386 noarch fat compatible os's : Linux CPUINFO: processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 7 model name : Pentium III (Katmai) stepping : 3 cpu MHz : 600.519 cache size : 512 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 mtrr pge mca cmov pat pse36 mmx fxsr sse up bogomips : 1202.13
*** Bug 215813 has been marked as a duplicate of this bug. ***
I've hit this on three machines now; IBM ThinkStation dual-P4, ThinkPad X31 P4 1.6, and a homebrew P4 2.8, all upgrades from FC5. In my case the solution has been to boot the rescue CD, or the DVD, or binary CD #1 in rescue mode, then # chroot # rpm -e kernel # rpm -i kernel-*.i686.rpm (from the install RPMS on the CD) This involves removing a few other packages (gnome-session, pcmciautil, etc.) for dependencies, but you can get either get them back from the install RPMS with the kernel, or with yum, if you don't mind doing it in text mode. I think this is bad enough to justify a re-issue of the CD/DVD images. Anyone like the poster above who has to do a mass upgrade with kickstart must be cursing Fedora right now.
First, you should have done a --nodeps kernel uninstall. Second, according to this page http://fedoraproject.org/wiki/Bugs/FC6Common the best way to go is to inatall yumdownloader and then do the following: yumdownloader kernel-2.6.18-1.2849.fc6.i686 su -c "rpm -Uvh --replacefiles --replacepkgs kernel-2.6.18-1.2849.fc6.i686.rpm" In the above I've updated the kernel version number to the latest one. You may need to do the same (twice!) if you follow the procedure after the next version comes out.
I'm seeing this too, rpm --showrc | head (without /etc/rpm/platform): ARCHITECTURE AND OS: build arch : i386 compatible build archs: pentium4 pentium3 i686 i586 i486 i386 noarch fat build os : Linux compatible build os's : Linux install arch : pentium4 install os : Linux compatible archs : pentium4 pentium3 i686 i586 i486 i386 noarch fat compatible os's : Linux /proc/cpuinfo: processor : 0 vendor_id : GenuineIntel cpu family : 15 model : 2 model name : Intel(R) Pentium(R) 4 CPU 2.66GHz stepping : 9 cpu MHz : 2660.442 cache size : 512 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe up cid xtpr bogomips : 5326.22 This is on a Fujitsu Siemens Scenic N desktop pc, with a D1562 baseboard aka a YBER082530 model.
If people experiencing this have an opportunity to reinstall could you try with the following updates.img http://people.redhat.com/pnasrat/211941-updates.img Instructions here: http://fedoraproject.org/wiki/Anaconda/Updates
I can report success with the update from #72 I have tried it on 7 machines with all the same specs and the same kickstart file and the right kernel was installed on all of them. Great job!
I tried this on one of the machines of the type which was causing trouble here and I too can confirm that this update fixes this. One other remark though, I did a ftp install with: "linux updates=http://people.redhat.com/pnasrat/211941-updates.img" and after downloading stage 2 through ftp it reasked me which interface to use and automaticly did dhcp again for that interface to download the updates image. I think this might be unwanted let me know if you want me to file a bug for this. BTW being able to specify an URL for updates= is great!
*** Bug 217182 has been marked as a duplicate of this bug. ***
*** Bug 212857 has been marked as a duplicate of this bug. ***
(In reply to comment #73) > I can report success with the update from #72 > I have tried it on 7 machines with all the same specs and the same > kickstart file and the right kernel was installed on all of them. > > Great job! I's wonderful!!! But shoud be better an update of the FC6 ISO release. In this period I'm organizing a group of meeting on linux and hacking wireless and I'm suggesting to choose FC6 (instead of Ubuntu...) but the problem of i586 kernel can be too big for a neophyte (like write the new Anaconda img on a floppy or pendrive by Windows). Bye Andrea Tassi
Updated a Compaq Evo Pentium 4 machine from FC5 to FC6. This gave me an i586 kernel with the matching i686 kernel-devel. A virgin install on a Fujitsu Lifebook has done the same. This gives real headaches when attempting to build additional kernel modules. Whatever is choosing the architecture isn't being consistent.
I have an update on this issue: I have reinstalled FC6 on the same machine about 4 times, half of those times it chose the wrong kernel. No changes to the hardware were performed. All reinstalls were performed within a week. Currently, I have a proper kernel. I never used the updates disk. Sadly, this makes this issue have no common pattern :(
(In reply to comment #29) > I successfully used > > rpm -Uvh --replacefiles --replacepkgs kernel-2.6.18-1.2798.fc6.i686.rpm > > to replace the kernel. The rpm man page says that --force is equivalent to all > of --replacepkgs, --replacefiles, and --oldpackage, and since this isn't an > older package, it should be equivalent to --force. I've confirmation from a tester that the above rpm command works. I also note that the same command is documented at http://fedoraproject.org/wiki/Bugs/FC6Common While annoying (hard to walk through on phone or IRC) I hope to see a patch to Anaconda soon. I have not experienced this bug on Servers running RHEL5 Beta2, which is a fork of FC6. Then again, I'm using the x86_64 arch. Hmm.
(In reply to comment #63) > I found a solution, you need to start Fedora Core 6 install with this command: > > linux i686 > > this instructs installer to use i686 kernel. I can confirm that booting with "linux i686" the FC6 install DVD (in a Intel Centrino) without using the updated Anaconda img the problem of i586 kernel remains. Andrea Tassi
I was having this problem (i586 kernel being installed). I was also having the cpuspeed problem. Booting from DVD with linux i686 fixed the i586 kernel problem but did _not_ fix the cpuspeed problem. I modified the original FC-6-i386-DVD.iso to add an images/updates.img (from the downloaded 211941-updates.img mentioned above): growisofs -M FC-6-i386-DVD.iso -R -graft-points\ images/updates.img=211941-updates.img Of course, the resultant ISO image has a different checksum. checking this at boot time fails with this error message: Unable to read disk checksum from the primary file descriptor. This probably means the disk was created without adding a checksum. Anyone know how to "add a checksum"? Anyway, I verfied the sha1sum on the original download and the sha1sum's of the modified image and the burnt DVD compared well. Using the growisofs command to add the update seems to be the least invasive way of doing this so I feel comfortable that the resultant image is good. I installed from the modified DVD and it also fixed the i586 kernel problem but did _not_ fix the cpuspeed problem. To fix the cpuspeed problem, I changed /etc/cpuspeed.conf to specify the following for DRIVER: DRIVER=p4-clockmod and restarted cpuspeed /etc/rc.d/init.d/cpuspeed restart No errors! No errors on reboot either! Does this make sense?
Precedure in Comment #24 worked beautifully for me as well. I was able to succesfully re-install all removed dependecies and my nvidia drivers from the livna repo as well without problems. I did lose the FC6 kernel entry from my /boot/grub/grub.conf, so reboot failed. I then booted using the FC6 resuce CD and restore my grub.conf from a backup that I had previously made. Jason's fix corrected my i686 header problem which then allowed me to compile my VMWare Workstation 5.5.3 vmmon module succesfully. Thank you Jason!
(In reply to comment #83) > > Of course, the resultant ISO image has a different checksum. > checking this at boot time fails with this error message: > > Unable to read disk checksum from the primary file descriptor. > This probably means the disk was created without adding a checksum. > > Anyone know how to "add a checksum"? Haven't tried it yet but this should work: (cd $RPMS; rpm -i --nodeps anaconda-runtime*) /usr/lib/anaconda-runtime/implantisomd5 --help /usr/lib/anaconda-runtime/implantisomd5 $ISO where $RPMS is the directory where this Fedora Core package is and $ISO is the name of the ISO image to implant the (md5) checksum in.
I have the same problem with CPU frequency. I updated kernel from i586 to i686 as it was recommended. But it did not help. The actual problem is that CPU frequency driver (/lib/modules/2.6.18-1.2868.fc6/kernel/arch/i386/kernel/cpu/cpufreq/acpi-cpufreq.ko) is not loaded. When I load driver manually (/sbin/insmod) everything getting ok. But how can cause to load the driver automatically even after kernel (and path to the module) update?
So is there any way to cause acpi-cpufreq.ko to load automatically?
Alexy, This is a general RH issue, and the most frequent solution I use is 1) make sure there is a reference in modprobe.conf if appropriate, 2) make sure the module is loaded, 3) force the initrd to be remade, using mkinitrd or similar tool. Ultimately, you can always resort to the rc.local file to get stuff loaded at boot.
*** Bug 221849 has been marked as a duplicate of this bug. ***
The "linux i686" option does not appear to work in upgrade mode. It leaves the i586 kernel. It might work on a fresh install... (In reply to comment #63) > I found a solution, you need to start Fedora Core 6 install with this command: > > linux i686 > > this instructs installer to use i686 kernel.
*** Bug 223413 has been marked as a duplicate of this bug. ***
Consider using Kernel Fix Script (kernel-fix.sh) which is available on FC6 Common Bugs and Issuse wiki page[1] if your system already have i586 kernel. This script has been tested on a fresh new install with help with RahulSundaram. [1] http://fedoraproject.org/wiki/Bugs/FC6Common
This still happens for a fresh install from "FC devel" with the latest "2.6.19" kernel based "rescue" image as of 2006-01-20. The affected system is an "IBM ThinkPad T23". The issue becomes apparent after the first start-up of the new system, because "CPU Frequency Scaling Monitor" claims that frequency scaling is not supported on this system. The content of "cpuinfo" after installing the "i686" kernel reads: processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 11 model name : Mobile Intel(R) Pentium(R) III CPU - M 1200MHz stepping : 4 cpu MHz : 798.000 cache size : 512 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 mtrr pge mca cmov pat pse36 mmx fxsr sse up bogomips : 1599.72 clflush size : 32
I was trying to help one user with this problem. He reported that "linux i686" did not work but "linux kernel=i686" did work. http://forums.fedoraforum.org/forum/showthread.php?t=144887
*** Bug 223802 has been marked as a duplicate of this bug. ***
*** Bug 224531 has been marked as a duplicate of this bug. ***
*** Bug 220995 has been marked as a duplicate of this bug. ***
same problem here but can't install i686 kernel # arch i686 processor : 0 vendor_id : AuthenticAMD cpu family : 6 model : 10 model name : AMD Athlon(tm) stepping : 0 cpu MHz : 1100.138 cache size : 512 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow ts bogomips : 2201.87 removed the i586 kernel and tried to install the latest i686 kernel: rpm -i kernel-2.6.19-1.2288.2.1.fc5.i686.rpm package kernel-2.6.19-1.2288.2.1.fc5 is intended for a i686 architecture
Same issue but on AMD`s ARCHITECTURE AND OS: build arch : i386 compatible build archs: athlon i686 i586 i486 i386 noarch fat build os : Linux compatible build os's : Linux install arch : athlon install os : Linux compatible archs : athlon i686 i586 i486 i386 noarch fat compatible os's : Linux processor : 0 vendor_id : AuthenticAMD cpu family : 6 model : 10 model name : AMD Athlon(tm) XP 2500+ stepping : 0 cpu MHz : 1826.176 cache size : 512 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow up ts bogomips : 3654.12 clflush size : 32 will there be a update for this in time ?
Created attachment 150461 [details] Updated fc6 kernel fix script.
Sorry first time with bugzilla. The above attachment is explained in this post. If one has already updated the kernel using the normal yum process then you may have two kernels installed on your system (Yum keeps two version so you can rollback if things break). This will break the script on the common issues page as the rpm queries return i586i586 instead of i586. ORing the i586 with a i586i586 takes care of this scenario. Also if one has extra RPM's of kernel drivers installed (Digium Zaptel 1.2 for Asterisk 1.2 ) then these will of course not work as post install, yum seems to work fine even if you have a i586 kernel so you get i686 hardware drivers installed. Also to untangle one self from wierd RPM conflict issues, now that the upgrades repo is filling up, it is sometimes beneficial to not upgrade the kernel but just move the existing kernel to i686. The above script will work on any kernel as it matches on the installed kernel version strings not the latest one in the repos and just sideways-grades (hmmm... don't think that will catch on but you get my meaning) to i686. Once you have rebooted you can yum update as normal and get the latest kernel. I have also included the fix for when you have two kernels installed as well. The attached script could probably do with some sanity checking and cleanup but I think it should probably replace the one on the common issues page in some form or other as it should work in more (most?) situations. In a related issue the install directive 'linux i686' does indeed not work as stated in an earlier message. It might be an idea to update the common issues page with 'linux kernel=i686'. Cheers
I am sure there is a better way of testing for the latest version of the installed kernel than using the i586i586 kludge above because obviously this won't work for people who have three or more versions of the kernel (I guess this is not normal anyway). I couldn't find any way to use RPM's -q option to output just the latest installed version. If there were a way to get RPM to spit out only the latest version then this would be a better way to do things in the script.
I'm kickstarting a lot of FC6 on ~50 machines during school classes. The anaconda is picking i586 kernel at a random choice. I see no reasons why repeated installs gives various results.
...I'm inserting Attansic 1Gbps driver in %post so I know that this is random even on the same machine even just reinstalling without HW change.
Are all your machines actually i686 class? If so, I've had pretty good luck using the suggestion to put "i686" on the boot line. (i.e. in the isolinux.cfg file on the CD) Once I made that change, it started consistently installing the i686 kernel for me.
Yes, they are all AM2 CPU. But I don't care about the kernel as the machines are in the classroom and they are reinstalled all the time we are starting new lesson and I compiled both i586 and i686 drivers and the %post script pick up the right version. To print out what arch version you are using run the command: rpm -qa --qf "%{NAME}-%{VERSION}-%{RELEASE}.%{ARCH}\n" kernel\*
I see, to use just i686 it needs to be in the isolinux.cfg file and starting the install from the commandline you need to use linux kernel=i686? I have tried to use linux i686 from the commandline and it did not work. I am sure this is reproducible as I run various test machines in a windows vmware environment. Milan, it might be ok for now but if your environment changes and you need to use kernel drivers from the repositories you might get unhooked as these will install in an expected i686 kernel tree which will not be present. Worth just bearing in mind. Also, the scipt uses an RPM command very similar to the above to determine the kernel version. The issue is that yum keeps the previous version of the kernel so you get a double entry returned which the script on the common issues wiki page does not handle in its if statement. The script is a pretty easy way resolve this issue as you don't have to have any technical knowledge (apart from the generic instructions for downloaded scripts). 1) download script 2) chmod +x <script name> 3) ./<script name> However this issue plus some others encouraged me to look at smart/smart-gui as a replacement for yum/yumex and to be honest I don't think I will go back to yum/yumex. After install I used smart GUI to look for and fix already installed packages (it found a conflicting oppenssl library). Then it updated my system. The biggest difference was in it's speed. It doesn't always try and update the local cache, You tell it to do it once and it uses this until you tell it to do it again (or after some timeout value). I know yum does this (it has a timeout) but yumex is an absolute dog to run especially with the extras repo. yum also will sit there for many seconds doing nothing if it hits a dodgy mirror (seems to be quite common these days). smart makes multiple connections so a bad connection does not hold everything up. I think smart is probably the best package manager for 64 bit Fedora (multilib environment) but I still have to prove this:) Anyway I digress... Rehan
Should this problem be fixed through a yum update in some way? i.e. machines that have the incorrect kernel architecture get it replaced with the correct one.
okay well i dont know how it happened but i was playing around with the kernels (upgrading i586 to i686) and destoyed the OS. I the put the os files back on (repair) and then the OS was completely ruined hahaha. So i loaded the fresh copy and still had the problems. So i tried: "rpm -Uvh --replacefiles --replacepkgs kernel-2.6.18-1.2798.fc6.i686.rpm" & "rpm -Fvh kernel-2.6.18-1.2798.fc6.i686.rpm". They didnt work. I read #29 stating "The rpm man page says that --force is equivalent to all of --replacepkgs, --replacefiles, and --oldpackage, and since this isn't an older package, it should be equivalent to --force". The RPM man is wrong. I then used: "rpm -Uvh --force kernel-2.6.18-1.2798.fc6.i686.rpm" and it worked; [root@localhost coza]# rpm -Uvh --force kernel-2.6.18-1.2798.fc6.i686.rpm Preparing... ########################################### [100%] 1:kernel ########################################### [100%] [root@localhost coza]# rpm -q --qf='%{name}-%{version}-%{release}.%{arch}\n' kernel kernel-2.6.18-1.2798.fc6.i686 this is to anyone who have tried serveral methods to no avail.
*** Bug 234026 has been marked as a duplicate of this bug. ***
*** Bug 218965 has been marked as a duplicate of this bug. ***
> [root@localhost coza]# rpm -Uvh --force kernel-2.6.18-1.2798.fc6.i686.rpm Better to use "rpm -ivh ..." to install a new kernel rather than upgrading the existing one. A failure could leave you with a non-bootable system with no fallback kernel.
But you can't install both the i586 and the i686 version of the same kernel without causing conflicts which can be worse than the -Uvh. What you can do now that there's newer kernels in the updates is to -ivh a _different_ kernel in its i686 version, which sidesteps the entire --force issue.
None of this seems like a reasonable solution. Surely the best solution is to sort this all out through yum updates. What worries me is that I can't find any sort of discussion of such a fix above. If it's not possible to do this with yum and rpm, then they obviously need modifying.
The updates.img contains the correct fix for this installing with that updates.img should be sufficient.
I was attempting to install the Nvidia Linux drivers using the instructions at http://www.nvnews.net/vbulletin/showthread.php?s=fbc9a10b5ef6b9b08cb447c9c13e7a32&t=72490. Following “ERROR: Unable to load the kernel module [...]”, I did yum -d0 list kernel <Enter> Installed Packages kernel.i586 2.6.19-1.2911.fc6 installed kernel.i586 2.6.19-1.2948.fc6 installed Available Packages kernel.i686 2.6.20-1.2948.fc6 updates I tried rpm from comment 110 but error in finding directory was reported. Jason Ferrell's method worked. yum -d0 list kernel kernel-devel <Enter> Installed Packages kernel.i686 2.6.19-1.2948.fc6 installed kernel-devel.i685 2.6.19-1.2948.fc6 installed Available Packages kernel.i586 2.6.19-1.2948.fc6 updated kernel-devel.i586 2.6.19-1.2948.fc6 updated On reboot, the boot loader pointed to the old 2.6.19-1.2911 and reported error 15, “File not found”. Using the editor offerred by Fedora at error, I replaced the old with the new 2.6.20-1.2948.fc6, and booted, <b>. I replaced old with new in /boot/grub/grub.conf. (In reply to comment #24) > I also ran into this problem on an AMD64 X2 3800+. It was causing strangeness > when attempting to install livna's nvidia drivers (arch i586 vs i686). > > My fix was to swap out the i586 kernel for the i686 with yum, rather than rpm > directly, and it worked just fine: > 1) yum remove kernel kernel-devel > 2) yum install kernel.i686 kernel-devel.i686 > 3) yum install (the list of a dozen odd i386 kernel depencies removed in step > #1, such as "gnome-volume-manager" and "compiz") >
A simpler and more intuitive method to fix this is to to use Smart Package Manager instead of yum to fix things. Do this: yum install smart fedora-package-config-smart add smart-gui if you want a kde/gnome frontend. then do: smart update (if it asks questions say YES) then do (to see what is installed: smart query installed kernel if you think the wrong kernel is installed you can then do: smart query kernel to get a list of possible kernels and (for example) smart install kernel-2.6.20-1.2948.fc6@i686 Smart Package Manager will fix all the things neccessary to get this kernel installed. Much better than yum/yum utils on the whole. It should also be very easy to modify script provided on the common issues wiki page to use Smart. Actually Smart is faster/more efficient/better day to day for package management allowing you to relegate yum to special circumstances where you might need super fine grain control over packages. The only issue with it is that it sucks to get the list of mirrors populated for your repositories so you end up hammering the main server. I have a scripted system which uses some generic flat files to build repository configuration files for both yum and smart (and ensure that there is a full list of mirrors for smart) and this will be found here: http://sourceforge.net/projects/pkgrautoupdater/ Just need to find some time to get the docs written and the files posted.