"plymouthd: Could not start boot splash, no such file or directory"

[NuLogicSystems]

When booting the live ISO with plymouth, if you hit esc or a function key plymouth is supposed to switch between itself and the text console. however this only works if you hit this twice really fast, after about a sec or two this no longer works and it can't find the splash any longer. This works on shutdown with no issues.

[NuLogicSystems]

I was watching the buildiso process and shouldn't plymouth be loaded after these in initcpio:
[miso_overlayfs]
[miso_loop_mnt]
[miso_pxe_common]
[miso_pxe_http]
[miso_kms]

Just a thought, I maybe wrong.

[NuLogicSystems]

gohlip posted this in the forum on another topic, but i'm thinking that maybe it's this kind of little mistake that could be part of the issue:
"Manjaro's grub puts the initrd line wrongly for separate boot partition.
There's initrd /boot/intel-ucode.img /boot/initramfs-xxxx.img
when it should be initrd /intel-ucode.img /initramfs-xxxx.img"

[philmmanjaro]

Well, the wiki says:

Add plymouth to the HOOKS array in mkinitcpio.conf. It must be added after base and udev for it to work:

/etc/mkinitcpio.conf
HOOKS="base udev plymouth [...] "

[philmmanjaro]

Seems I was totally confused by your post @AJSlye. I restored the old grub-2.02.beta2-10 as else Arch never would have adopted our patch.

[philmmanjaro]

ucode is always installed at /boot/intel-ucode.img. make_system_path_relative_to_its_root /boot/intel-ucode.img detects the relative path and creates /intel-ucode.img if needed.

[philmmanjaro]

@AJSlye: did you test to re-arrange the mkinitcpio.conf to your suggestion and tested it yet?

[NuLogicSystems]

No, I went to bed last night.

I'm going to take a look at the 2014.09 ISO which plymouth works correctly on, and see what might have changed since then. The 2014.09 was based on the Manjaro-kde-0.8.10 ISO.

After searching the net I found a couple of possible reasons for this issue, one has to do with kms and the other an fstab issue.

[philmmanjaro]

There is no issue what so ever. Otherwise you won't even been able to boot your installation after selecting a separate /boot partition. This would have been noticed by our community from the get-go. I also tested the current version of grub. No issues to been found:

[johnd@john-pc Desktop]$ cat /etc/fstab
# /etc/fstab: static file system information.
#
# Use 'blkid' to print the universally unique identifier for a device; this may
# be used with UUID= as a more robust way to name devices that works even if
# disks are added and removed. See fstab(5).
#
# <file system>                           <mount point>  <type>  <options>  <dump>  <pass>
UUID=c0ab3fd8-c1bc-41ae-89ab-834f4783f188 /boot          ext4    defaults,noatime 0       2
UUID=6c78034b-e53e-4c53-adba-0ba3b12a0f2a /              ext4    defaults,noatime 0       1
[johnd@john-pc Desktop]$ sudo update-grub
[sudo] password for johnd: 
Generating grub configuration file ...
Found background: /usr/share/grub/background.png
Found Intel Microcode image: /boot/intel-ucode.img
Found linux image: /boot/vmlinuz-4.1-x86_64
Found initrd image: /boot/initramfs-4.1-x86_64.img
Found initrd fallback image: /boot/initramfs-4.1-x86_64-fallback.img
Found memtest86+ image: /boot/memtest86+/memtest.bin
done
[johnd@john-pc Desktop]$ sudo cat /boot/grub/grub.cfg | grep intel-ucode
    initrd  /intel-ucode.img /initramfs-4.1-x86_64.img
        initrd  /intel-ucode.img /initramfs-4.1-x86_64.img
        initrd  /intel-ucode.img /initramfs-4.1-x86_64-fallback.img

[NuLogicSystems]

I wasn't referring to the intel-ucode itself, just the possibility that something might be pointed in the wrong direction. While researching this issue, I found that this error can come up when plymouthd can't reconnect to the graphics adaptor and/or find the proper resolution again (KMS / vesa mode setting). The other reason this error can occur had to do with the fstab, but I can't seem find that forum post again. There was also a post about the systemd service file for plymouth and an AFTER= tag.

[philmmanjaro]

Well, plymouth is not officially supported by Arch. If we find a solution I'm happy for it. However current grub might not be related to this issue here.

[NuLogicSystems]

Agreed, however if it is in the grub file I'll let you know, but not until
after I have a look at whatever changes leszek had made to your
Manjaro-kde-0.8.10 released ISO for netrunner-rolling 2014.09 that made
this work. This is not a show stopper bug for me, just an annoyance. You
can go ahead and either close it, move it elsewhere or whatever else you
think is appropriate.

[philmmanjaro]

Simply keep me updated if you find something.

[NuLogicSystems]

No problem.

[NuLogicSystems]

Using the latest ISO that I built, if I hit esc I still get the virtual console with the output in it, however now when I hit esc the second time I no longer get this error, it's just saying starting plymouth splash again, but the plymouth splash still isn't coming back up. I'm wondering now if this is actually an issue caused by one of the plymouth systemd service / unit files as suggested by one of the threads I had read earlier.

[NuLogicSystems]

OK, your version of Plymouth is 0.8.9-1.
In the AUR there are 3 pkgbuilds:
plymouth 0.9.2-3
plymouth-git 0.9.2.r0.g2c437c3-1
and
plymouth-legacy 0.8.8-4

I'm not sure which one of these your pkgbuild is based on or if your using your own custom pkgbuild, but I'm going to do some tests with these pkgbuilds this weekend, in between other things, to see if any of these would fix this issue.

[NuLogicSystems]

I'm also confused by one of your post, is grub-2.02.beta2-10 older than upstreams grub-1:2.02.beta2-5?

[philmmanjaro]

grub-2.02.beta2-10 is from the source code newer than that what Arch is shipping. I released a repacked version as grub-2.02.beta2-11 as you got me confused with the ucode stuff. Now the old package is restored.

[philmmanjaro]

I've now added a new package called plymouth-dev. This was tested by Eugen F. Also we had a regression with 0.8.9 in regard of luks so the normal plymouth package is back to 0.8.8.

[NuLogicSystems]

Does the new plymouth-dev package fix the issue here?

[philmmanjaro]

You may want to try it yourself if your issue is fixed by this. Eugen just tests it against password entries with luks encryption.

[NuLogicSystems]

I finally got a chance to try this out and still no luck.