View previous topic :: View next topic |
Author |
Message |
crazzyfool n00b
Joined: 13 Jan 2009 Posts: 37
|
Posted: Sun May 12, 2013 3:54 pm Post subject: Gentoo on GPT + Software RAID + LVM with GRUB 2 on non-EFI s |
|
|
Hi,
My installation just failed on the reboot. It gets to the Grub2 boot menu and I'm able to select my kernel. It begins to load but fails due to the following error:
Code: |
[0.419920] kvm: VM_EXIT_LOAD_IA32_PERF_GLOBAL_CRTL does not work properly. Using workaround
...
[0.445706] io scheduler noop registered
|
I based my installation on a combination of the official documents and the below website. My setup is very similar to the one in the below link.
http://kirneh.eu/gentoo-on-gpt-software-raid-lvm-with-grub-2-on-non-efi-system/
The only real difference is that I am using two SSDs in a RAID-0 / LVM volume. And two SATA-II disks, again in a RAID-0 / LVM configuration.
Does anyone know what the problem might be? Or what (and how) I can check from the LiveCD?
I have included my /etc/fstab in case this helps:
Code: |
# <fs> <mountpoint> <type> <opts> <dump/pass>
/dev/mapper/vg--ssd-lv--boot /boot ext4 noauto,noatime,nodiratime,discard 1 2
/dev/mapper/vg--ssd-lv--root / ext4 noatime,nodiratime,discard 0 1
/dev/mapper/vg--hdd-lv--home /home ext4 noatime 0 2
/dev/mapper/vg--hdd-lv--var /var ext4 noatime 0 2
/dev/mapper/vg--hdd-lv--usrportage /usr/portage ext4 noatime 0 2
/dev/mapper/vg--hdd-lv--vhdd /hdd ext4 noatime 0 2
/dev/sdc1 none swap sw,pri=1 0 0
/dev/sdd1 none swap sw,pri=1 0 0
tmpfs /tmp tmpfs noatime,nodiratime,size=5G 0 0
/dev/sr0 /mnt/dvd auto noauto,ro,user 0 0
/dev/sr1 /mnt/dvdrw auto noauto,ro,user 0 0
|
Cheers,
Andy |
|
Back to top |
|
|
BradN Advocate
Joined: 19 Apr 2002 Posts: 2391 Location: Wisconsin (USA)
|
Posted: Sun May 12, 2013 6:37 pm Post subject: |
|
|
Those errors don't look fatal. Can you take a picture of the last output visible?
Also, how did you build the kernel, genkernel, custom config, ?
The last time I did a manual config, I must have missed something important and the system locked up at boot. Genkernel worked though, so I started with one of its configurations and then altered it to how I wanted it. |
|
Back to top |
|
|
crazzyfool n00b
Joined: 13 Jan 2009 Posts: 37
|
Posted: Mon May 13, 2013 7:48 pm Post subject: |
|
|
Thanks for the help Brad. I have emailed you the picture.
Yeah I compiled a custom kernel but did use genkernel to create the initramfs:
Code: | genkernel --lvm --mdadm --install initramfs |
Not sure what I could have missed?
Do you think I could have something to do with the SSDs? |
|
Back to top |
|
|
BradN Advocate
Joined: 19 Apr 2002 Posts: 2391 Location: Wisconsin (USA)
|
Posted: Mon May 13, 2013 8:00 pm Post subject: |
|
|
I can't find the email, what subject did you use?
Maybe something to try would be taking the config used on the livecd and modifying that, there are so many options in the kernel these days that trial and error could take a long time. |
|
Back to top |
|
|
crazzyfool n00b
Joined: 13 Jan 2009 Posts: 37
|
|
Back to top |
|
|
BradN Advocate
Joined: 19 Apr 2002 Posts: 2391 Location: Wisconsin (USA)
|
Posted: Sun May 19, 2013 7:50 pm Post subject: |
|
|
Well, a few tips to keep in mind:
You can manually copy the .config file in/out of the kernel source directory, use that to manage your own versions of it.
After you change the .config file, it's good to do "make oldconfig", especially if it came from a different kernel version.
genkernel has an option --oldconfig to make it *not* generate its own configuration and use the one you supplied, when you are having it compile a whole kernel.
So basically, back up your current .config, try out genkernel (letting it make its own configuration, kernel, and initrd), then try out the resulting stuffs and see if it boots successfully. If so, great, use the .config it came up with as a basis for your own modifications, then either build manually like you were, or use genkernel --oldconfig to let it do the work (use "all" target and whatever other options you need, esp. for the initrd options if you use raid/lvm/etc).
For example, on my grub2 system, all it takes to make a new kernel and have it added to the grub boot menu is:
#(set up the .config file as desired in the kernel source directory)
#ln the desired kernel source to /usr/src/linux, as this is what genkernel uses
genkernel --oldconfig --mdadm --lvm --makeopts=-j4 all
grub2-mkconfig > /boot/grub.cfg
These last two lines I put in a shell script to save typing.
The other idea I suggested was starting with a gentoo livecd kernel config that boots on your system, which you can do by copying /proc/config.gz from a system running that kernel. Same principles, un-gzip that config.gz to the kernel source tree of your choosing (maybe you want to try a kernel version similar to that used on the livecd if you can't get newer ones to boot), make oldconfig, and build as per above.
Good luck! |
|
Back to top |
|
|
crazzyfool n00b
Joined: 13 Jan 2009 Posts: 37
|
Posted: Mon Jun 03, 2013 10:33 pm Post subject: |
|
|
BradN - thanks man! Good advice!
I recompiled the kernel letting genkernel do all the work. To a mixed feeling of joy and despair - it worked! That prompted me to start the installation from scratch. This time I was a little more discerning about what modules to include - which envolved a lot of greping of the working .config file!
It worked!........kind of....
The first thing I did was check the internet settings but found that ifconfig only had the lo interface. After trying to bring eth0 up I get:
Code: |
# /etc/init.d/net.eth0 start
* Bringing up interface eth0
* ERROR: interface eth0 does not exist
* Ensure that you have loaded the correct kernel module for your hardware
* ERROR: net.eth0 failed to start
|
Obviously now I'm thinking that I've missed another module but when I check the logs I see:
Code: |
[0.904331] r8169 Gigabit Ethernet driver 2.3LK-NAPI loaded
[0.904609] r8169 0000:0a:00.0: irq 69 for MSI/MSI-X
[0.904739] r8169 0000:0a:00.0: eth0: RTL8168c/8111c at 0xffffc9000181e000, 00:24:8c:7b:5f:0b, XID 1c4000c0 IRQ 69
[0.904902] r8169 0000:0a:00.0: eth0: jumbo features [frames: 6128 bytes, tx checksumming: ko]
[0.905065] r8169 Gigabit Ethernet driver 2.3LK-NAPI loaded
[0.905330] r8169 0000:09:00.0: irq 70 for MSI/MSI-X
[0.905435] r8169 0000:09:00.0: eth1: RTL8168c/8111c at 0xffffc9000181c000, 00:24:8c:7b:60:cf, XID 1c4000c0 IRQ 70
[0.905597] r8169 0000:09:00.0: eth1: jumbo features [frames: 6128 bytes, tx checksumming: ko]
|
....and then toward the end I see:
Code: |
[7.416698] systemd-udevd[7203]: starting version 200
[7.448024] systemd-udevd[7217]: renamed network interface eth0 to enp10s0
[7.457027] systemd-udevd[7219]: renamed network interface eth1 to enp9s0
|
Any idea what's going
Thanks. |
|
Back to top |
|
|
crazzyfool n00b
Joined: 13 Jan 2009 Posts: 37
|
Posted: Mon Jun 03, 2013 10:45 pm Post subject: |
|
|
Lol - I just came across the below thread!
https://forums.gentoo.org/viewtopic-t-961048.html
All I had to do was:
touch /etc/udev/rules.d/80-net-name-slot.rules
...and reboot - solved!
Thanks again! |
|
Back to top |
|
|
BradN Advocate
Joined: 19 Apr 2002 Posts: 2391 Location: Wisconsin (USA)
|
Posted: Mon Jun 03, 2013 11:05 pm Post subject: |
|
|
Yeah the interface renaming "feature" is a source of some headaches sometimes. It's really exciting on remote machines where if your interface doesn't come up, you're in some amount of trouble.
Back in the old days configuring your own kernel from hand wasn't so bad - not nearly as much option overload. But, now there's just so much stuff! I always used to do all configuration from a default .config but the last time I had to give up and start from something that worked - I still don't know what option was messing it up. |
|
Back to top |
|
|
crazzyfool n00b
Joined: 13 Jan 2009 Posts: 37
|
Posted: Sat Jun 08, 2013 1:02 pm Post subject: |
|
|
Looks like I might have been a bit premature with my celebration as I still appear to have a couple of issues with my configuration!
So, the system boots and I can’t see anything too sinister in the logs. However, every time I reboot I see messages like the one below:
Code: |
* Failed to shutdown Logical Volume Manager
* Wiping internal cache failed
* Error mdraid failed to stop
* mdadm: fail to stop array /dev/md0: Device or resource busy. Perhaps a running process, mounted filesystem or active volume group?
|
I’m not sure if this is a problem or not? The md0 device has / and /boot on it. md1 seems to unmounts fine when rebooting. If this is not an issue – can it be removed as an error when rebooting? Or maybe try and unmount it slightly later when rebooting?
The other issue I’m having, which could be unrelated, is that I can’t mount my nfs (which I’ve done many times from the LiveDVD during the installation).
Code: |
storm ~ # mount -t nfs "192.168.111.250:/personal shared" /root/nfs/ -o sync
mount: wrong fs type, bad option, bad superblock on 192.168.111.250:/personal shared,
missing codepage or helper program, or other error
(for several filesystems (e.g. nfs, cifs) you might
need a /sbin/mount.<type> helper program)
In some cases useful info is found in syslog - try
dmesg | tail or so
storm ~ #
|
Any ideas? Any help greatly appreciated!
Cheers,
Andy |
|
Back to top |
|
|
crazzyfool n00b
Joined: 13 Jan 2009 Posts: 37
|
Posted: Mon Jun 10, 2013 3:15 pm Post subject: |
|
|
Any ideas/ thoughts / suggestions?
Cheers,
Andy |
|
Back to top |
|
|
BradN Advocate
Joined: 19 Apr 2002 Posts: 2391 Location: Wisconsin (USA)
|
Posted: Tue Jun 11, 2013 6:10 pm Post subject: |
|
|
I think this may be normal when running / on raid/lvm. As long as you don't get raid de-sync problem after rebooting, I wouldn't worry too much. |
|
Back to top |
|
|
crazzyfool n00b
Joined: 13 Jan 2009 Posts: 37
|
Posted: Tue Jun 11, 2013 6:29 pm Post subject: |
|
|
Thanks for your reply - I really appreciate it!
Is there a way I can remove the error so I don't see it each time my computer goes down for a reboot?
Cheers,
Andy |
|
Back to top |
|
|
BradN Advocate
Joined: 19 Apr 2002 Posts: 2391 Location: Wisconsin (USA)
|
Posted: Tue Jun 11, 2013 6:37 pm Post subject: |
|
|
Probably it could be done by editing init scripts in /etc/init.d. Personally I would just leave it, or cover it with fbsplash. |
|
Back to top |
|
|
|