[SOLVED] Fresh install, Two failures and hang

15goudreau · 2014-06-20 00:22:22

Solution: Needed to mount my boot partition (EF filesystem) to /boot/efi instead of /boot. Using rEFInd

Hi, so I've read the beginners guide inside and out and I'm having the most difficult time installing arch.

I've probably wiped it at least a dozen times and everything seems like it is going to start up and I get a few failures on startup which results in a hang afterwards

1)

[FAILED] Failed to start Create Volatile Files and Directories.
see 'systemctl status systemd-tmpfiles-setup.service' for details.

2) (this one cycles 6 times before hang)

[FAILED] Failed to start login service.
See 'systemctl statuse systemd-logind.service' for details.

x6
Then I get

[OK] Reached Target Multi-User System.
[OK] Reached target graphical interface.

then the hang where I can't do anything but hold down my power button.

I tried restarting the system with my liveCD (usb) and mounting up my directories and looking into the systemd files but they just say they load ok and there aren't any issues.

I've scoured google and really can't find anything to fix this issue. Any advice would be welcome for sure!

I know I can install Arch on this laptop because I was successful once. But then I wiped it to try out OpenSUSE, and since then I can't get passed these errors.

Thanks in advance!

Last edited by 15goudreau (2014-06-28 16:45:53)

mcloaked · 2014-06-20 07:36:15

You are not likely to get any sensible replies unless you provide more complete information about exactly what you did for your install. Mentioning that you read the wiki does not say what you actually did, and since there are going to be a lot of variations on how you installed your system that might give a clue as to why the system is giving problems. It also matters that different hardware (which machine, which graphics card, etc) will affect the outcome. You also did not say whether you are using legacy boot or UEFI. So provide some details and you might have a chance of getting some help.

Last edited by mcloaked (2014-06-20 07:36:38)

15goudreau · 2014-06-20 10:09:58

Thanks for replying. I absolutely understand.

I'm installing with UEFI on my laptop. Specifically Asus A52F. The graphics are Intel HD integrated. I formatted my drive for GPT with 5 partitions sda1 > root, sda2 > home, sda3 > boot, sda4 > var, sda5 > etc.
My boot was EF00 while everything else was default.
I formatted with ext4 for all but the boot which was fat 32.
I installed gummiboot as my bootloader, and arch as my hostname.
I was installing over Wifi not ethernet, which seemed to work fine.
Any other details I might be missing? If so, I would be happy to provide them!

also in order to boot from the flashdrive I had to change the bootx64.efi to the loader.efi else I would get a blue box error.

Last edited by 15goudreau (2014-06-20 12:26:09)

15goudreau · 2014-06-20 12:23:56

So I figured I would paste the systemctl info here
after I remount all my partitions to /mnt with the live CD here are the outputs

systemd-tmpfiles-setup.service - create volatile files and directories
  loaded: loaded (/usr/lib/systemd/system/systemd-tmpfiles-setup.service; static)

  active: active (exited) since fri 2014-06-20 12:06:09 utc; 1 min 36s ago

  docs: man:tmpfiles.d(5)

          man:systemd-tmpfiles(8)
process: 214 ExecStart=/usr/bin/systemd-tmpfiles --create --remove --boot --exclude-prefix=/dev (code=exited, status=0/SUCCESS)
main PID: 213 (code=exited, status=0/SUCESS)
CGroup: /system.slicee/systemd-tmpfiles-setup.service
Jun 20 12:06:09 archiso systemd[1]: started create volatile files and directories

and

systemd-logind.service - login service
loaded: loaded (usr/lib/systemd/system/systemd-logind.service; static)
active: active (running) since FRI 2014-06-20 12:06:10 UTC; 7min ago
docs: man:systemd-logind.service(8)
    man:logind.conf(5)

    [url]http://www.freedesktop.org/wiki/software/systemd/logind[/url]

    [url]http://www.freedesktop.org/wiki/software/systemd/multiseat[/url]
main PID: 239 (systemd-logind)
status: "processing requests..."
CGroup: /system.slice/systemd-logind-service
  ->239 /usr/lib/systemd/systemd-logind
Jun 20 12:06:10 archiso systemd-logind[239]: watching system buttons on /dev/input/event2 (power button)
Jun 20 12:06:10 archiso systemd-logind[239]: watching system buttons on /dev/input/event4 (video bus)
Jun 20 12:06:10 archiso systemd-logind[239]: watching system buttons on /dev/input/event0 (lid switch)
Jun 20 12:06:10 archiso systemd-logind[239]: watching system buttons on /dev/input/event1 (sleep button)
Jun 20 12:06:10 archiso systemd[1]: started login service
Jun 20 12:06:10 archiso systemd-logind[239]: watching system buttons on /dev/input/event0 (lid switch)
Jun 20 12:06:10 archiso systemd-logind[239]: watching system buttons on /dev/input/event2 (power button)
Jun 20 12:06:10 archiso systemd-logind[239]: watching system buttons on /dev/input/event1 (sleep button)
Jun 20 12:06:10 archiso systemd-logind[239]: watching system buttons on /dev/input/event4 (video bus)
Jun 20 12:06:15 archiso systemd-logind[239]: new session c1 of user root.

mcloaked · 2014-06-20 12:46:16

Immediately after you completed your install were you able to get a console login without graphical.target being enabled? Did you use pacstrap to do a base and base-devel install or something different to this. During the install, once you had booted to the usbkey, did you mount the root partition, boot partition (and maybe the others)? Once you did that did you arch-chroot into your newly created and populated partitions on the hard drive, and then what did you install, if anything, before rebooting to your new system?

What systemd units are enabled ahead of the reboot? Did you create a non-root user when you were still in the chroot before rebooting? There may well be other questions that need answering before anyone can offer any real help, but this would be a start.

Did you try to go straight into a graphical login or did you try to login as multi-user first to check that all was well before setting up a graphical login for a D.E.? Also which install iso did you use to do the install? Was it the June 2014 archiso?

Also since you are using gummiboot for your bootloader did you read the wiki at https://wiki.archlinux.org/index.php/Gummiboot?

Last edited by mcloaked (2014-06-20 12:49:48)

15goudreau · 2014-06-20 13:03:14

Q1) No, immediately after install I wasn't able to get a console login. I get those errors with the hang and can't type anything. I did use pacstrap to do a base and base-devel install.

pacstrap -i /mnt base basedevel

yes to install

I mounted all the partitions I created, which were /, /home, /boot, /var, /etc. all to the path of /mnt

mount /dev/sda1 /mnt
mkdir /mnt/home
mkdir /mnt/boot
mkdir /mnt/var
mkdir /mnt/etc
mount /dev/sda2 /mnt/home
mount /dev/sda3 /mnt/boot
mount /dev/sda4 /mnt/var
mount /dev/sda5 /mnt/etc

After

arch-chroot /mnt /bin/bash

I generated the locale of en_US.UTF-8
I exported LANG=en_US.UTF-8
I edited my vconsole with

KEYMAP=dvorak
FONT=lat9w-16

then timezone

ln -s /usr/share/zoneinfo/US/Eastern /etc/localtime

then hardware clock

hwclock --systohc --UTC

I added my hostname arch to hosts
I installed: iw wpa_supplicant and dialog
I created the initial ram disk environment with

mkinitcpio -p linux

I then installed gummiboot with

pacman -S gummiboot
gummiboot install

I edited the arch.conf for boot as

title           Arch Linux
linux         /vmlinuz-linux
initrd        /initramfs-linux.img
options    root=/dev/sda1 rw

then I exited the chroot, umounted the /mnt, and rebooted. Upon reboot those failures.

I'm not sure how to see which systemd units are enabled ahead of reboot.
I did not create a non-root user while in chroot
couldn't get a graphical login because of hang,
I installed from June 2014 ISO.

Only thing I see different between my gummiboot install on the beginners guide and the gummiboot wiki is under options I use /sda1 where they use a PARTUUID.
would that make a difference? I'm certain my root partition is sda1

I did find out when I hang after reaching target graphical interface, if I press ctrl+alt+delete the system processes that keystroke as reboot and will reboot my system. I don't know if that helps.

Last edited by 15goudreau (2014-06-20 13:11:39)

Head_on_a_Stick · 2014-06-20 17:48:13

Did you edit /boot/loader/loader.conf & change the "default" entry to "arch" (your gummiboot .conf file)?

Last edited by Head_on_a_Stick (2014-06-20 17:49:06)

mcloaked · 2014-06-20 19:04:12

You can check the status of any systemd unit with the systemctl status command - for example in my case:

# systemctl status graphical.target
● graphical.target - Graphical Interface
   Loaded: loaded (/usr/lib/systemd/system/graphical.target; static)
   Active: active since Wed 2014-06-18 17:51:26 BST; 2 days ago
     Docs: man:systemd.special(7)

Jun 18 17:51:26 home1 systemd[1]: Starting Graphical Interface.
Jun 18 17:51:26 home1 systemd[1]: Reached target Graphical Interface.

You can interrupt gummiboot during boot (see https://wiki.archlinux.org/index.php/ke … Gummiboot) and change the kernel line to add systemd.unit=multi-user which should make the system boot to a non-graphical console and give a simple command line prompt for login as root. If that is the case then you can at least check that the system does boot and then explore what is wrong with the graphical login once you have a command line login.

Also you did not mention installing a desktop environment or window manager? I don't know how the system would respond if it was attempting to boot to a graphical login without one. In my case I install kde which includes the kdm graphical login greeter and the status for me shows it is enabled:

# systemctl status kdm
● kdm.service - K Display Manager
   Loaded: loaded (/usr/lib/systemd/system/kdm.service; enabled)
   Active: active (running) since Wed 2014-06-18 17:51:17 BST; 2 days ago
 Main PID: 319 (kdm)
   CGroup: /system.slice/kdm.service
           ├─319 /usr/bin/kdm -nodaemon
           └─327 /usr/bin/X :0 vt7 -nolisten tcp -auth /var/run/xauth/A:0-VmMiHb

Jun 18 17:51:17 home1 systemd[1]: Starting K Display Manager...
Jun 18 17:51:17 home1 systemd[1]: Started K Display Manager.
Jun 18 17:51:23 home1 kdm[335]: :0[335]: pam_unix(kde:session): session opened for user mi...d=0)
Hint: Some lines were ellipsized, use -l to show in full.

It is also worth checking during the install that you were booted to the install media in uefi mode - you can check that by doing:

efivar -l

which should give a reasonable number of lines of output similar to:

a4c5067a-5b95-4770-9ac0-59e6544ca239-WOL
8be4df61-93ca-11d2-aa0d-00e098032b8c-BootCurrent
8be4df61-93ca-11d2-aa0d-00e098032b8c-LangCodes
8be4df61-93ca-11d2-aa0d-00e098032b8c-PlatformLangCodes
8be4df61-93ca-11d2-aa0d-00e098032b8c-ErrOut
8be4df61-93ca-11d2-aa0d-00e098032b8c-ErrOutDev
8be4df61-93ca-11d2-aa0d-00e098032b8c-BootOptionSupport
8be4df61-93ca-11d2-aa0d-00e098032b8c-ConInDev
560bf58a-1e0d-4d7e-953f-2980a261e031-PNP0510_0_VV
560bf58a-1e0d-4d7e-953f-2980a261e031-PNP0501_1_VV
560bf58a-1e0d-4d7e-953f-2980a261e031-PNP0501_0_VV
560bf58a-1e0d-4d7e-953f-2980a261e031-PNP0400_0_VV
560bf58a-1e0d-4d7e-953f-2980a261e031-PNP0604_0_VV
8be4df61-93ca-11d2-aa0d-00e098032b8c-ConOutDev
ec87d643-eba4-4bb5-a1e5-3f3e36b20da9-UsbMassDevValid
ec87d643-eba4-4bb5-a1e5-3f3e36b20da9-UsbMassDevNum
0885f288-418c-4be1-a6af-8bad61da08fe-DriverHlthEnable
7459a7d4-6533-4480-bba7-79e25a4443c9-DriverHealthCount
8be4df61-93ca-11d2-aa0d-00e098032b8c-OsIndicationsSupported
97ca1a5b-b760-4d1f-a54b-d19092032c90-DebuggerSerialPortsEnabledVar
560bf58a-1e0d-4d7e-953f-2980a261e031-SerialPortsEnabledVar
etc....

Are you sure that your gummiboot install command worked properly during the install? It should have created nvram entries in the motherboard boot memory - you can check what there is by installing efibootmgr if you didn't already do so, and then use efibootmgr on its own as root - which will list your nvram entries and show which is default

eg in my case I have:

# efibootmgr
BootCurrent: 0000
Timeout: 1 seconds
BootOrder: 0000,0002,0001,0003,0004,000A
Boot0000* Arch Linux (rEFInd)
Boot0001* SATA : PORT 6G 0 : M4-CT256M4SSD2 : PART 0 : Boot Drive
Boot0002* SATA : PORT 5 : M4-CT032M4SSD3 : PART 0 : Boot Drive
Boot0003* LAN : IBA GE Slot 00C8 v1376
Boot0004* LAN : IBA GE Slot 0200 v1376
Boot000A* UEFI : SATA : PORT 5 : M4-CT032M4SSD3 : PART 0 : OS Bootloader

I use rEFInd and not gummiboot (by the way there is a small chance that your system could be affected by a known uefi boot bug which can affect particular hardware for individual kernels at random leading to a kernel that won't boot but that is a small chance of happening).

Anyway it is worth exploring that you have set up your gummiboot config files correctly, and also check the nvram entries as outlined above.

15goudreau · 2014-06-20 19:05:29

Right now in loader.conf it says
#timeout 3
default 052495b09c8d41f18e70cb1ee93c3d7d-*
are you saying it should just say default arch?

15goudreau · 2014-06-20 19:10:45

I am booting into UEFI I know that for sure. I did not install any DE or WM during the install, as I was going to do that after when I had my root terminal in startup. I don't know if gummiboot installed properly, however I did try earlier with GRUB and it experienced the same issue. I will try to interupt gummiboot and insert that command so that I can at least get to the terminal. Hopefully I will be able to check the journal as well.

I wouldn't think that my computer is affected by that bug as I was able to get a working install on my very first try a month ago with gummiboot, but maybe something has changed. Thank you guys for your help so far, hopefully we can get to the bottom of this. I will try those suggestions and report back

15goudreau · 2014-06-20 20:39:37

Alright,
I added the systemd.unit=multi-user in gummiboot and then booted here is what happened

Still see the

[failed] to start create volatile files and directories.
other stuff...
[ok] reached target system initialization
       starting rescue shell...
[ok] started rescue shell
[ok] reached target rescue mode.
welcome to rescue mode! type "systemctl default" or ^D to enter default mode.
type "journalctl -xb" to view system logs. type "systemctl reboot" to reboot.
sulogin: cannot open password database.

And then cursor hang where again, nothing works.

I can't grab the systemctl status graphical.target because the only way to access that is not in chroot. But then it just pulls the information from my liveCD.

Is there anyway that my keyboard isn't being recognized during the bootup process, so what seems like a hang, is just no communication? The only thing that would be odd about that is that ctrl+alt+delete causes the system to reboot fine.

If I do another fresh install I could install efimanager and check to make sure it is configured properly.

Last edited by 15goudreau (2014-06-20 20:44:27)

mcloaked · 2014-06-20 21:07:53

Did you also install util-linux when you installed the system? I found an old thread https://bbs.archlinux.org/viewtopic.php?pid=1166116. I don't know if it is related to your error output concerning [failed] to start etc. But just a long shot.

One other question - how did you create your usbkey to do the install? Did you use the dd method? From your output I think that somewhere along the way during your install you may have missed a step. I have not had any installs in the past couple of years fail to boot after the initial install. However it is of course vital that particularly the bootloader or bootmanager is installed and configured correctly. Also it might be an idea to post your /etc/fstab contents - just to make sure that it looks sane. Also I note that you said that you had wiped your original arch install to try OpenSuse - did you reformat your partitions after they were used for Suse? If not then it might also be a good idea to clean up the partitions before installing arch on them.

I also note that you originally mentioned that your partitions were "ext4 for all but the boot which was fat 32". The ESP has to be Fat 32 as well of course. If the EFI system partition is not Fat32 then you won't get a successful UEFI boot. If your /boot is the ESP and is Fat32 then that will work fine.

What is the output of the following two commands when you are in the chroot:

# df -h
# ls -l /boot

You presumably have a directory within /boot called /boot/EFI which is the EFI System Partition? See https://wiki.archlinux.org/index.php/Un … _Interface. In which directory is your efi binary for gummiboot installed?

Is your hard drive formatted with a partition table that is GPT or legacy MBR?

Last edited by mcloaked (2014-06-21 08:27:39)

Head_on_a_Stick · 2014-06-21 09:08:17

15goudreau wrote:

Right now in loader.conf it says
#timeout 3
default 052495b09c8d41f18e70cb1ee93c3d7d-*
are you saying it should just say default arch?

Yes, your loader.conf should read:

timeout 3
default arch

15goudreau · 2014-06-22 16:46:07

@ mcloaked:

I don't believe I installed util-linux, as a separate package. Unless it was included in base or base-devel, I didn't install it. I used a few programs to create the USB key as I also thought it might be the cause. I used unetbootin-windows application. And I tried universal_usb_installer. Both with the may version and june version of the .iso. I think I reformatted the drive? I used the --zap-all command before each install. And when I used cgdisk, it always told me there was an issue with the disk and it would try to fix it. When I verified it with cgdisk, it said the disk was fine for GPT. I'm a bit confused at your statement about the ESP. I have a partition called /boot which is formatted with EF00 for the efi portion. I was under the assumption I only needed one partition named /boot where I was to install my bootloader. I think the latter portion of your statement is correct, my /boot is the ESP which is formatted FAT 32.

Contents of my fstab

#
# /etc/fstab: static file system information
#
# <file system> <dir>     <type>    <options>      <dump>    <pass>
# /dev/sda1
UUID=0ce47657-a08e-4991-b1fe-3daeb987c159       /                    ext4                        rw,relatime,data/ordered      0     1

#   /dev/sda2
UUID=7359ff31-85bb-4f31-98d9-033da0dfb3f5             /home         ext4                         rw,relatime,data=ordered      0      2

#   /dev/sda4
UUID=3f00956a-f5e5-46e8-905b-85792b8bf6dd         /var               ext4                         rw,relatime,data=ordered      0      2

#   /dev/sda5
UUID=0b31ceae-f127c-4aad-8844-796a1c0abdef       /etc            ext4                           rw,relatime,data=ordered      0       2

#   /dev/sda3
UUID=4f9B-a44a                 /boot               vfat              rw,relatime,fmast=0022,dmast=0022,codepage=437,iocharset=iso8859-1,shortname=mixed,errors=remount-ro        0         2

output of df -h is

df : []/run/archiso/bootmnt[]: No such file or directory
df: []/run/archiso/cowsapace[]: no such file or directory
df: []/run/archiso/sfs/root-image[]: no such file or directory
df: []/sys/fs/cgroup/systemd[]: no such file or directory
df: []/sys/fs/cgroup/cpuset[]: no such file or directory
df: []/sys/fs/cgroup/cpu,cpuacct[]: no such file or directory
df: []/sys/fs/cgroup/memory[]: no such file or directory
df: []/sys/fs/cgroup/devices[]: no such file or directory
df: []/sys/fs/cgroup/freezer[]: no such file or directory
df: []/sys/fs/cgroup/net_cls[]: no such file or directory
df: []/sys/fs/cgroup/blkio[]: no such file or directory
df: []/run/user/0[]: no such file or directory
Filesystem                                  Size       Used     Avail       Use%        mounted on
/dev/mapper/arch_root-image         15g        661M     14G        5%           /
dev                                              2.9G      0            2.9G         0%         /dev
run                                              2.9G       0            2.9G       0%          /run
tmpfs                                           2.9G        0          2.9G        0%          /dev/shm
tmfps                                           2.9G         0          2.9G       0%           /tmp
hugetlbfs                                      2.9G        0           2.9G       0%           /dev/hugepages
/dev/sda5                                    9.8G          32M      9.2G        1%          /etc
/dev/sda2                                    15G           38M       14G         1%         /home
/dev/sda3                                    511M         24M       488M      5%         /boot
/dev/sda4                                    14G           227M      13G        2%         /var
udev                                            2.9G          0              2.9G      0%        /dev
shm                                            2.9G          0               2.9G     0%        /dev/shm
run                                             2.9G           0             2.9G        0%       /run
tmp                                             2.9G           0           2.9G        0%        /tmp
/dev/mapper/arch_root-image         1.5G           686M      745M     48%        /etc/resolv.conf

and
ls -l /boot

total 23728
drwxr-xr-x 4 root root     4096 Jun 20 01:10 EFI
-rwxr-xr-x  1 root root     17102304 Jun 20 01:09 initramfs-linux-fallback.img
-rwxr-xr-x 1 root root     3265622 Jun 20 01:09 initramfs-linux-img
drwxr-xr-x 3 root root          4096 Jun 22 01:55 loader
-rwxr-xr-x 1 root root       3914848 Jun 17 07:34 vmlinuz-linux

I'm going to try the cygin method next to install on a flashdrive and see if I can get around these errors.

@ head_on_a_stick
When I change my loader.conf to default arch I get the same set of errors.

15goudreau · 2014-06-22 16:50:42

When I chrooted into my system and ran pacman -S util-linux it said

1/1 reinstalling util-linux

So I would guess it was installed

Edit:

Tried installing LXDM and xorg in chroot. Still got errors, so I unintalled with pacman -Rsn. Also something else I noticed, might be trivial: when boot with my liveCD it says "welcome to Arch Linux!", when I boot normally it says "Welcome to Linux". Not sure if that helps at all.

Last edited by 15goudreau (2014-06-22 17:06:42)

mcloaked · 2014-06-22 17:36:43

The standard method for creating the usb key is to use the dd command from linux - using the command:

# dd bs=4M if=/path/to/archlinux.iso of=/dev/sdx && sync

See https://wiki.archlinux.org/index.php/US … tion_Media

That method produces a special hybrid partition table on the key that allows both legacy bios boot and uefi boot if the system has uefi boot available, and does not need the usb key to be reformatted beforehand as it re-writes the partition table as well as populating the necessary directories. I seem to remember that using unetbootin will not produce a usbkey that will correctly boot with uefi. It is very important that when the usbkey is booted for the install that you check at that stage that it has booted using uefi and not legacy bios by looking at the output of the efivar -l command. If the efi vars are not loaded then when you write the boot stuff it will not correctly write the uefi boot parameters unless you are already booted in uefi mode.

When formatting your hard drive on the computer that you will be installing on, the best method when using uefi is to use the gdisk command (it is provided within the install media and you can format the HDD during the initial stages of the install process). gdisk was written by the author of the refind boot manager especially to format drives using a gpt partition table and it does so efficiently and without error in my experience. See https://wiki.archlinux.org/index.php/GU … tion_Table

From the output that you posted your /boot with the directory /boot/EFI/ looks fine, and I presume that within that directory is your gummiboot directory and the gummiboot.efi file is in that directory? The fact that your /boot partition is EF00 means that you do have a GPT partition table. Did you check that the partition has a boot flag? You can use parted to look at the partitions on the drive and check that there is a boot flag on the ESP. Without it there is a chance that the boot process won't recognise the ESP to boot from it.

It would be nice to see the output from the commands:

ls -l /boot/

ls -l /boot/EFI/

The reason is that your kernel and initramfs files should be in /boot by default, and the gummiboot directory in /boot/EFI/

It is possible that the method you have used to prepare your install usbkey is not booting to uefi mode for the install - if you could verify the output from the efivar -l command with the key booted that would at least tell us if you really are booted in the correct mode for the install. If for any reason you were not booted to that mode, but in legacy bios boot, then I think that gummiboot installing the boot files would not do so correctly.

It is best to verify this before trying to diagnose further.

Last edited by mcloaked (2014-06-22 17:47:02)

15goudreau · 2014-06-22 18:16:13

When I boot off my USB I press escape during the POST. I have two options to select to boot from that have to do with my USB key. one is sandisk U3 titanium 3.27, the other is EUFI: sandisk u3 titanium. Anyways, when I select the UEFI option it brings me to another screen that has no arch linux splash screen (like the legacy does). And the options there are Arch Linux Archiso x86_64 UEFI USB, UEFI shell x86_64 v1, UEFI shell x86_64 v2. After I boot up off the drive It says "welcome to "arch linux!" other stuff and then the normal root login.

Here is the output from efivar -l

8be4df61-93ca-11d2-aa0d-00e098032b8c-IgdVBIOSRevH
ef152fb4-7b2f-427d-bdb4-7e0a5826e64-Bootflow
numbers-PlatformLangCodes
numbers-LangCodes
numbers-ConErrDev
numbers-ASUSCheckKeyPressed
numbers-ConInDev
numbers-ConOutDev
numbers-UsbMassDevNum
numebrs-SerialPortsEnabledVar
numbers-SBRev
-NBRev
-SetupVolatileData
-boot0004
-loaderentrydefault
-boot0000
-bootorder
-timeout
-cpus3resume
-defaultlegacydevorder
-defaultbootorder
-platformlang
-lang
-LDP
-ConIn
-ConOut
-AEDID
-MemCeil.
-VerbTabLeS3Save
-PCISerialPortsLocationVar
-PchInit
-Setup
-Setupcpufeatures
-ExtdAcpiGlobalVariable
-AcpiGlobalVariable
-MonotonicCounter
-boot0003
-legacyDevOrder
-AMITSESetup
-OEM3
-SLP2
-Boot0002
-LegacyDevChecksum
-VKEY
-CNFG
-MFGO
-STDefaults

ls -l /boot/ > output I posted in my previous post
ls -l /boot/EFI/

Total 8
drwxr-xr-x 2 root root 4096 June 20 01:10 Boot
drwxr-xr-x 2 root root 4096 Jun 20 01:10 Gummiboot

Parted gives me the following when I run "print"

Number          Start      End          Size    File        system       name      flags
1                  1049kb   16.1 GB  16.1GB  ext4
2                  16.1Gb    32.2 GB   16.1 GB  ext4
3                 32.2GB    32.8GB     537Mb    fat32                                   boot
4                 32.8GB    47.8GB     15.0GB   ext4      
5                47.8GB      58.5GB     10.7GB   ext4

I tried using dd in windows as I do not have a computer that has linux on it. When I try and run the command "dd if=image.iso of=/dev/sdb bs=4M" which for me is dd if=archlinux-2014.06.01-dual.iso of=/dev/sdb bs=4M" I get an error that says it cannot find the output file.

I tried to download and install cygwin, and it wouldn't install properly during the download from the mirrors, I have absolutely no idea why.

Edit: Reinstalled with gdisk instead of cgdisk and I got the exact same error as before.

Last edited by 15goudreau (2014-06-22 18:48:36)

mcloaked · 2014-06-22 19:26:00

OK this is a puzzle! I seem to remember that there is no arch splash screen when booting to uefi in the install media so that sounds correct. Your /boot contents looks OK, and the kernel and initrd files are in the right place. One other thing worth checking is the output from the efibootmgr command (you may have to pacman -S efibootmgr in the chroot if you don't have the package) - that should tell you if the nvram is set up correctly for booting to the gummiboot entry. It is just possible that you may be unlucky and that your particular machine is one where the creation of nvram entries leads to a situation where despite writing the nvram entry it does not show up in the firmware menu. See the troubleshooting section at https://wiki.archlinux.org/index.php/Bo … leshooting - but first it is worth checking the output of the efibootmgr command as that should at least confirm if the nvram has an entry. The other possibility is that occasionally the nvram entry is present but not written correctly. It is then necessary to explore if the kernel can be loaded by booting the uefi shell and then loading the kernel from there, or even by adding another entry to boot the kernel efistub directly. See https://wiki.archlinux.org/index.php/EF … UEFI_Shell. I had a system a year ago where I had to resort to writing the nvram entry using the bcfg command after booting to the uefi shell v2 - see https://wiki.archlinux.org/index.php/Un … rface#bcfg

If these alternative methods fail to boot the kernel then I am puzzled as to where the problem lies. Another route is to try a different bootloader such as grub instead of gummiboot - this will not use the efistub but will boot the kernel via a different method that does not rely on the kernel efistub. So this will need more work to identify what the problem is - but I can understand your frustration. It should be straightforward to get the system booted!

PS - is your BIOS version 218? you can check using

dmidecode|grep BIOS

in the chroot (you may need to install dmidecode first). Also the latest BIOS for your machine was about two and a half years ago from what I can tell - and there have been a lot of developments in the UEFI boot firmware in the BIOS of many machines. It is just a possibility that the BIOS in your machine may be buggy.

Last edited by mcloaked (2014-06-22 20:06:30)

mcloaked · 2014-06-22 19:48:20

There is still a possibility that you may be unlucky and have come across the bug that has a long history - see https://bugs.archlinux.org/task/33745. There are also upstream bugs reports such as the kernel bug where someone was able to boot using the uefi shell - see the bug comment at https://bugzilla.kernel.org/show_bug.cgi?id=68761#c22

The other suggestion may be to simply try one of the bootloaders that does not reply on the efi stub loader - such as grub or sysylinux.

15goudreau · 2014-06-22 20:11:47

I just looked at the section you mentioned. In regards to the bootx64.efi. I mentioned this earlier but I want to highlight it again. In order to get my thumbdrive to even boot, I had to basically remove the bootx64.efi from the EFI portion and substitute it with the loader.efi renamed to bootx64.efi. I'm not sure if that helps in anyway.

The process of debugging isn't so frustrating, what is frustrating is that it worked my first time ever setting it up and now it refuses to boot after I try a different distro. But sometimes computer are just weird. I really appreciate your help! I'm going to do that debug and report back.-

Edit: saw your other post. My bios is 218. And I updated it before I started this process I think. Because I was having issues booting. However that didn't help me any. I don't think there is an update for my laptop past 218. I'll check with ASUS. Maybe a downgrade is needed?

Last edited by 15goudreau (2014-06-22 20:13:44)

15goudreau · 2014-06-22 20:23:55

Ok dmidecode Says

SMBIOS entry point at 0xbae01198
SMBIOS 2.6 present.
BIOS information
           BIOS is upgradeable
           BIOS Shadowing is allowed
           Bios Boot specification is supported
     BIOS revision: 4.6
Bios Language information

efibootmgr displays

BootCurrent: 0004
timeout: 2 seconds
bootorder: 0001,0000,0002,0003,0004
boot0000* linux boot manager
boot0001* linux boot manager
boot0002* CD/DVD Drive
boot0003* Hard drive
boot0004* UEFI:Sandisk U3 Titanium 3.27

Edit: Because there are two linux boot managers that prompted me to explore them in my options. The first one 0000 returns an error saying a bootable device needs to be inserted. The second one 0001 loads the boot screen but displays the error. My harddrive just displays a blinking cursor that doesn't do anything. Almost as though after the errors I get through the linux boot manager, it passes on to my hard drive and it can't do anything....

ran the create microsoft directory code and such. Rebooted. No change. I'm going to run the other debug things you mentioned.

Tried booting from UEFI shell. When I ran cd \EFI\arch I get "target directory not found" and \EFI\archlinux also gets target directory not found

Tried doing it directly without boot manager

efiboot mgr -d /dev/sda3 -p 1 -c -L "Arch Linux" -l /vmlinuz-linux -u "root=/dev/sda1 rw initrd=/initramfs-linux.img"
**warning ** :Boot0005 has same label Arch Linux

******************************************************************
Warning! This MBR disk does not have a unique signature.
If this is not the first disk found by EFI, you may not be able to boot from it without a unique signature.
Run efibootmgr with the -w flag to write a unique signature to the disk.
*********************************************************************

******************************************************************
Warning! This MBR disk does not have a unique signature.
If this is not the first disk found by EFI, you may not be able to boot from it without a unique signature.
Run efibootmgr with the -w flag to write a unique signature to the disk.
*********************************************************************


bootcurrent:0004
timeout:2 seconds
bootorder: 0006,0005,0001,0000,0002,0003,0004
boot0000* linux boot manager
boot0001* linux boot manager
boot0002* cd/dvd drive
boot0003* hard drive
boot0004* UEFI:Sanddisk u3 titanium 3.27
boot0005* arch linux
boot0006* arch linux

I think there are two arch linux boot entries because I ran that code twice

I did try GRUB earlier but got the same results. Maybe it would just be easier to switch to GRUB? Can I install grub on top of Gummiboot and it will erase gummiboot? I have no preference to my bootloader really. Well, I take that back, one that works.

Last edited by 15goudreau (2014-06-22 21:02:20)

mcloaked · 2014-06-23 09:27:08

15goudreau wrote:

efibootmgr displays
BootCurrent: 0004
timeout: 2 seconds
bootorder: 0001,0000,0002,0003,0004
boot0000* linux boot manager
boot0001* linux boot manager
boot0002* CD/DVD Drive
boot0003* Hard drive
boot0004* UEFI:Sandisk U3 Titanium 3.27
Edit: Because there are two linux boot managers that prompted me to explore them in my options. The first one 0000 returns an error saying a bootable device needs to be inserted. The second one 0001 loads the boot screen but displays the error. My harddrive just displays a blinking cursor that doesn't do anything. Almost as though after the errors I get through the linux boot manager, it passes on to my hard drive and it can't do anything....
ran the create microsoft directory code and such. Rebooted. No change. I'm going to run the other debug things you mentioned.
Tried booting from UEFI shell. When I ran cd \EFI\arch I get "target directory not found" and \EFI\archlinux also gets target directory not found

OK it is not entirely clear what the boot process is doing here and it looks like the most likely problem is a gummiboot config error. Your boot order in the efibootmgr command shows that entry 0001 is the first one that it will try which you have as "linux boot manager" - which presumably will find the gummiboot entry but perhaps it is not configured correctly. Certainly you don't have any \EFI\arch or \EFI\archlinux directories since you previously quoted the contents of the /boot/EFI/ directory using the ls -l command. in \EFI\ there are only a Boot and a Gummiboot directory.

I am not clear why you had to run the command that created the boot entries twice but presumably the entry at boot0000 is likely invalid from the first attempt. You can use the efibootmgr command to delete nvram entries as well as recreate them. I expect it might be a good idea to delete entry 0000, and maybe also 0001 and then create a new entry yourself using the efibootmgr command.

When you say that entry 0001 "loads the boot screen" - do you mean that it does successfully start the gummiboot bootloader but that gummiboot then fails to load the kernel? If so then that would suggest that the problem lies with the gummiboot configuration files, but that the uefi boot process does start correctly and loads gummiboot. If that is the case then I would probably delete the nvram entry 0000 and then make sure that the default order was showing that the first entry that will be tried will be gummiboot (which seems to have its entry as "linux boot manager". I don't use gummiboot myself so I don't have knowledge of what its boot screen looks like.

You can use the efibootmgr command to write a new entry similar to :

efibootmgr -c -d /dev/sdX -p Y -l /EFI/Gummiboot/gummiboot.efi -L "Gummiboot"

where sdX is the drive you have for ESP, and Y is the partition number. So in your case sdX is sda and Y is 3 if your partitions are as you reported previously. Using this command will then give an entry in the nvram called Gummiboot rather than "linux boot manager". For the meaning of the parameters see the output of the command "man efibootmgr".

Check the man command for efibootmgr to see how to remove and reorder the boot entries in nvram. Make sure that the entries and boot order are sane before trying to reboot. Also make sure you only delete the entries that you created and not the ones that will boot to the usbkey or the CD.

If this does start Gummiboot correctly then you need to diagnose why gummiboot is not loading the arch kernel correctly. You can follow the wiki page at https://wiki.archlinux.org/index.php/Gummiboot and set up a manual entry to load the arch kernel as the file /boot/loader/entries/arch.conf containing:

title          Arch Linux
linux          /vmlinuz-linux
initrd         /initramfs-linux.img
options        root=PARTUUID=14420948-2cea-4de7-b042-40f67c618660 rw

Here the PARTUUID value to the right of the = sign is the value relating to your root partition which is in your case the output of

blkid -s PARTUUID -o value /dev/sda1

Perhaps this will help pinpoint where the problem lies. Certainly it is looking increasingly like your gummiboot config files are where the boot fails.

If you did want to try to launch your arch kernel from the uefi shell then boot to uefi shell v2 as you seem to have tried before. In your case you will need to tell the uefi shell where the kernel and initrd files are so presuming that your files system is defined as fs0 then from the shell in your case it would then be:

> fs0:
> cd \EFI\
> vmlinuz-linux root=PARTUUID=14420948-2cea-4de7-b042-40f67c618660 rootfstype=ext4 add_efi_memmap initrd=EFI/initramfs-linux-img

Since the kernel and initrd files in your case are named vmlinuz-linux and initramfs-linux-img respectively. Again the numerical value of the PARTUUID is as before the output of the blkid I quoted before.

You only need to change directory to \EFI\ since your kernel and initrd are in /boot in your case. All of these commands have to be carefully adapted to your specific case and can't just be cut and pasted from the wiki entry which only has sample commands in many cases.

I hope this helps.

Last edited by mcloaked (2014-06-23 16:53:35)

15goudreau · 2014-06-23 11:57:22

Wow, that is a very in depth reply. I really can't thank you enough for your time! I ran the line of code twice which created the entries because after I ran it the first time I rebooted to see if it worked, but it didn't so I ran it again so that you could see the output. Looking back I had the entry wrong with my sda set to sda3 and my partition number set to 1. I didn't bring my laptop to work, so when I get home I will delete all the entries which don't make sense and then rerun the manual entry with the correct configuration. Question regarding PARTUUID. Is there any difference specifying my root device as partuuid vs just uuid? I know there are two numbers, but I'm unsure of the difference. Since you said to use PARTUUID that's what I will use, I'm just curious.

In regards to the 0001 boot entry, I would assume it is loading the gummiboot and then the kernel, which fails. But since it would seem that my computer was never configured properly, I'm not sure I even know what the gummiboot loader looks like and I might have been using the linux boot manager instead. Hopefully when I change the entry, it looks different and loads properly, but I agree with you that it would seem like the gummiboot config files are affecting the failure.

Is there a reason as to why the gummiboot config files aren't working properly? Is it an issue which I created during the install process or does this sometimes happen? I'm just curious because eventually I want to migrate my desktop to Arch and dualboot with Windows, but I want to make sure I have some of these issues ironed out before hand, so that I'm not going to experience a lot of downtime. Obviously, if I am doing something wrong during the configuration of gummiboot, I want to know so I can learn from it! Else I'm just copying and pasting code in attempts to fix it and that isn't what I was trying achieve by switching to Arch.

Last edited by 15goudreau (2014-06-23 12:11:00)

mcloaked · 2014-06-23 13:26:13

OK 15goudreau, hopefully when you fix the nvram entry as well as getting the gummiboot stanza correct for using the appropriate parameters to tell it where the kernel and initrd files are then it should work. Concerning PARTUUID and UUID then yes they are two different strings - you can see both using a command such as the following on my system.

blkid  /dev/sdb1
/dev/sdb1: UUID="D8B9-6A05" TYPE="vfat" PARTUUID="08d86b83-7c6d-4e74-b45d-ce436f4fdced"

The command I sent previously just picks out the one value. Also you can run the blkid command on its own (as root) and it will list the values for all partitions.

Some boot managers allow the use of either provided you specify whether it is the PARTUUID or UUID but I don't know if gummiboot allows root=UUID=... instead of root=PARTUUID=.... though you can of course explore the man pages for gummiboot!

The information about booting via uefi shell and loading the kernel manually was simply another alternative possible way to check if your system was bootable but hopefully you will be able to get it working with gummiboot once the config files are fixed.

Of course you have a free choice as to whether you wish to use gummiboot, grub, rEFInd or another boot manager or bootloader - when I was making that decision a year or so back I decided on rEFInd, but it is a free world! There is some base information about gummiboot at http://freedesktop.org/wiki/Software/gummiboot/

Good luck and I hope that when you reconfigure your boot files it will give you a successful boot. For your question about whether you did something wrong when installing gummiboot I will have to leave to a gummiboot expert - I only use rEFInd (and grub) for uefi boot.

Last edited by mcloaked (2014-06-23 16:51:34)

15goudreau · 2014-06-24 11:57:06

Alright, so I did what you said to create the gummiboot boot order. And when it starts, all I get is a blinking cursor, nothing else. So that's a bummer. I'm going to try and load the kernel directly from the shell again and see what happens.

When I try to boot into shell v2 I get the following errors:

assert_efi_error (status = not found)
assert e:\dev\tips\edk2\mdemodulepkg\library\uefihiiserviceslib\uefihiiserviceslib.c(88): !efi-Error
(status)

assert_efi_error (status = not found)
assert e:\dev\tips\edk2\mdemodulepkg\library\uefihiiserviceslib\uefihiiserviceslib.c(94): !efi-Error
(status)

assert_efi_error (status = not found)
assert e:\dev\tips\edk2\mdemodulepkg\library\uefihiiserviceslib\uefihiiserviceslib.c(100): !efi-Error
(status)

I must have been trying to boot from the kernel in shell v1, which is why it didn't work.

So I tried installing GRUB and here is the error output after it loaded. I selected arch and the kernel loads and fails with hte same errors as before and a few new ones.

[failed] failed to start loginservice.
see 'systemctl status systemd-logind.service' for details.
[ok] reached target multi-user system.
[ok] reached target graphical interface.
[failed] failed to start rotate log files.
see'systemctl status logrotate.service' for details.
[failed] failed to start verify integrity of password and group files.
see 'systemctl status shadow.service' for details
[failed] failed to start update man-db cache.
see 'systemctl status man-db.service' for details

and then cursor hang, where pressing the power button once doesn't shutdown the computer ctrl + alt + delete reboots computer.

thoughts now that I'm using grub? I might try rEFI next....

so I re-ran grub again and now it isn't displaying the errors above, but it is still hanging where it usually does after [ok] reached target graphical interface.

So I have two USB drives so I'm thinking of trying some other ways of installing Arch which may be more successful. First is installing mint to a flashdrive and then while in mint, using DD to format the other flashdrive correctly knowing that I can now rule out the flashdrive being a source of error. Or the other way is using mint to install arch using the "use another version of linux to install arch" I don't know if this will fix the bootloader issues. But it might be worth trying?

Last edited by 15goudreau (2014-06-24 13:30:26)

Arch Linux

#1 2014-06-20 00:22:22

[SOLVED] Fresh install, Two failures and hang

#2 2014-06-20 07:36:15

Re: [SOLVED] Fresh install, Two failures and hang

#3 2014-06-20 10:09:58

Re: [SOLVED] Fresh install, Two failures and hang

#4 2014-06-20 12:23:56

Re: [SOLVED] Fresh install, Two failures and hang

#5 2014-06-20 12:46:16

Re: [SOLVED] Fresh install, Two failures and hang

#6 2014-06-20 13:03:14

Re: [SOLVED] Fresh install, Two failures and hang

#7 2014-06-20 17:48:13

Re: [SOLVED] Fresh install, Two failures and hang

#8 2014-06-20 19:04:12

Re: [SOLVED] Fresh install, Two failures and hang

#9 2014-06-20 19:05:29

Re: [SOLVED] Fresh install, Two failures and hang

#10 2014-06-20 19:10:45

Re: [SOLVED] Fresh install, Two failures and hang

#11 2014-06-20 20:39:37

Re: [SOLVED] Fresh install, Two failures and hang

#12 2014-06-20 21:07:53

Re: [SOLVED] Fresh install, Two failures and hang

#13 2014-06-21 09:08:17

Re: [SOLVED] Fresh install, Two failures and hang

#14 2014-06-22 16:46:07

Re: [SOLVED] Fresh install, Two failures and hang

#15 2014-06-22 16:50:42

Re: [SOLVED] Fresh install, Two failures and hang

#16 2014-06-22 17:36:43

Re: [SOLVED] Fresh install, Two failures and hang

#17 2014-06-22 18:16:13

Re: [SOLVED] Fresh install, Two failures and hang

#18 2014-06-22 19:26:00

Re: [SOLVED] Fresh install, Two failures and hang

#19 2014-06-22 19:48:20

Re: [SOLVED] Fresh install, Two failures and hang

#20 2014-06-22 20:11:47

Re: [SOLVED] Fresh install, Two failures and hang

#21 2014-06-22 20:23:55

Re: [SOLVED] Fresh install, Two failures and hang

#22 2014-06-23 09:27:08

Re: [SOLVED] Fresh install, Two failures and hang

#23 2014-06-23 11:57:22

Re: [SOLVED] Fresh install, Two failures and hang

#24 2014-06-23 13:26:13

Re: [SOLVED] Fresh install, Two failures and hang

#25 2014-06-24 11:57:06

Re: [SOLVED] Fresh install, Two failures and hang

Board footer