switch_root: failed to unlink ro_branch still present on rc3 #137

Closed
opened 2015-09-19 15:14:44 +00:00 by philm · 23 comments
philm commented 2015-09-19 15:14:44 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: NuLogicSystems

Can we please see if we can fix this?
Not that this message is critical, it's just annoying.

*Created by: NuLogicSystems* Can we please see if we can fix this? Not that this message is critical, it's just annoying.
philm commented 2015-09-19 16:17:17 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: philmmanjaro

Why, it doesn't hold us back to boot the image. If you find time to debug it, fine. At the moment it is low prio on my end.

*Created by: philmmanjaro* Why, it doesn't hold us back to boot the image. If you find time to debug it, fine. At the moment it is low prio on my end.
philm commented 2015-09-19 20:42:42 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: NuLogicSystems

I accidentally hit the wrong button, Clemens doesn't want to see this error.

*Created by: NuLogicSystems* I accidentally hit the wrong button, Clemens doesn't want to see this error.
philm commented 2015-09-19 21:25:21 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: philmmanjaro

Here is an really old take on the matter. Do you still use aufs on your end?

*Created by: philmmanjaro* Here is an really old [take on the matter](http://chakraos.org/forum/viewtopic.php?id=8492). Do you still use **aufs** on your end?
philm commented 2015-09-19 21:29:30 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: philmmanjaro

Also here the part of the aufs manual:

config AUFS_BR_RAMFS
    bool "Ramfs (initramfs/rootfs) as an aufs branch"
    help
    If you want to use ramfs as an aufs branch fs, then enable this
    option. Generally tmpfs is recommended.
    Aufs prohibited them to be a branch fs by default, because
    initramfs becomes unusable after switch_root or something
    generally. If you sets initramfs as an aufs branch and boot your
    system by switch_root, you will meet a problem easily since the
    files in initramfs may be inaccessible.
    Unless you are going to use ramfs as an aufs branch fs without
    switch_root or something, leave it N.

Current aufs config can be found here.

*Created by: philmmanjaro* Also here the part of the **aufs** manual: ``` config AUFS_BR_RAMFS bool "Ramfs (initramfs/rootfs) as an aufs branch" help If you want to use ramfs as an aufs branch fs, then enable this option. Generally tmpfs is recommended. Aufs prohibited them to be a branch fs by default, because initramfs becomes unusable after switch_root or something generally. If you sets initramfs as an aufs branch and boot your system by switch_root, you will meet a problem easily since the files in initramfs may be inaccessible. Unless you are going to use ramfs as an aufs branch fs without switch_root or something, leave it N. ``` Current **aufs config** can be found [here](https://github.com/manjaro/packages-core/blob/master/linux41/config.aufs).
philm commented 2015-09-19 21:41:30 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: philmmanjaro

However, we might even try the Ubuntu-way:

CONFIG_AUFS_BDEV_LOOP=y
# CONFIG_AUFS_BRANCH_MAX_1023 is not set
CONFIG_AUFS_BRANCH_MAX_127=y
# CONFIG_AUFS_BRANCH_MAX_32767 is not set
# CONFIG_AUFS_BRANCH_MAX_511 is not set
# CONFIG_AUFS_BR_FUSE is not set
CONFIG_AUFS_BR_HFSPLUS=y
# CONFIG_AUFS_BR_RAMFS is not set
# CONFIG_AUFS_DEBUG is not set
# CONFIG_AUFS_EXPORT is not set
# CONFIG_AUFS_FHSM is not set
CONFIG_AUFS_FS=m
# CONFIG_AUFS_HNOTIFY is not set
# CONFIG_AUFS_RDU is not set
CONFIG_AUFS_SBILIST=y
# CONFIG_AUFS_SHWH is not set
# CONFIG_AUFS_XATTR is not set
*Created by: philmmanjaro* However, we might even try the **Ubuntu-way**: ``` CONFIG_AUFS_BDEV_LOOP=y # CONFIG_AUFS_BRANCH_MAX_1023 is not set CONFIG_AUFS_BRANCH_MAX_127=y # CONFIG_AUFS_BRANCH_MAX_32767 is not set # CONFIG_AUFS_BRANCH_MAX_511 is not set # CONFIG_AUFS_BR_FUSE is not set CONFIG_AUFS_BR_HFSPLUS=y # CONFIG_AUFS_BR_RAMFS is not set # CONFIG_AUFS_DEBUG is not set # CONFIG_AUFS_EXPORT is not set # CONFIG_AUFS_FHSM is not set CONFIG_AUFS_FS=m # CONFIG_AUFS_HNOTIFY is not set # CONFIG_AUFS_RDU is not set CONFIG_AUFS_SBILIST=y # CONFIG_AUFS_SHWH is not set # CONFIG_AUFS_XATTR is not set ```
philm commented 2015-09-19 21:43:19 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: Kirek

the message is still there with overlayfs

*Created by: Kirek* the message is still there with overlayfs
philm commented 2015-09-19 21:48:04 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: philmmanjaro

We might need to take a look at our hooks as in:

    NOTES
           switch_root will fail to function if newroot is not the root of a mount. 
           If you want to switch root into a directory  that  does  not  meet this 
           requirement then you can first use a bind-mounting trick to turn any 
           directory into a mount point:

                  mount --bind $DIR $DIR
*Created by: philmmanjaro* We might need to take a look at our **hooks** as in: ``` NOTES switch_root will fail to function if newroot is not the root of a mount. If you want to switch root into a directory that does not meet this requirement then you can first use a bind-mounting trick to turn any directory into a mount point: mount --bind $DIR $DIR ```
philm commented 2015-09-19 21:58:12 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: NuLogicSystems

Yes I'm using overlayfs as well.

*Created by: NuLogicSystems* Yes I'm using overlayfs as well.
philm commented 2015-09-19 22:13:48 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: NuLogicSystems

This could also be what is causing the plymouthd issue:
If you sets initramfs as an aufs branch and boot your
system by switch_root, you will meet a problem easily since the
files in initramfs may be inaccessible.

*Created by: NuLogicSystems* This could also be what is causing the plymouthd issue: If you sets initramfs as an aufs branch and boot your system by switch_root, you will meet a problem easily since the files in initramfs may be inaccessible.
philm commented 2015-09-19 22:37:56 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: philmmanjaro

For more info on the hooks you can take also a look here

*Created by: philmmanjaro* For more info on the hooks you can take also a look [here](https://github.com/manjaro/manjaroiso/commits/master/src/hooks/miso)
philm commented 2015-09-19 22:42:54 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: NuLogicSystems

I'll have to look at that when I have time to sit and read it. ;)

*Created by: NuLogicSystems* I'll have to look at that when I have time to sit and read it. ;)
philm commented 2015-09-19 22:44:10 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: NuLogicSystems

On a side note: have you already looked at systemd 2.2.5's switch to take the place of su?

*Created by: NuLogicSystems* On a side note: have you already looked at systemd 2.2.5's switch to take the place of su?
philm commented 2015-09-19 22:48:09 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: Kirek

Its 225 :)
Yes, I looked at machinectl shell, though it doesn't replace the su binary or anything like that, even if it does almost the same thing.

*Created by: Kirek* Its 225 :) Yes, I looked at machinectl shell, though it doesn't replace the su binary or anything like that, even if it does almost the same thing.
philm commented 2015-09-19 22:55:56 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: NuLogicSystems

So it does almost the same thing but not the same thing. Hmm, interesting, I'll need to do some reading when I get a chance. :)

*Created by: NuLogicSystems* So it does almost the same thing but not the same thing. Hmm, interesting, I'll need to do some reading when I get a chance. :)
philm commented 2015-09-19 23:02:19 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: Kirek

Look are the original issue about that: https://goo.gl/ZGvjNM

*Created by: Kirek* Look are the original issue about that: https://goo.gl/ZGvjNM
philm commented 2015-09-20 00:31:03 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: philmmanjaro

Well, I searched a lot. switch_root apparently only works on true initrd (cpio style, not using squashfs). So one option would be to silence switch_root in the init hook on livecd-kernel-image. Archlinux don't has it as they use dmsnapshot.

*Created by: philmmanjaro* Well, I searched a lot. **switch_root** apparently only works on true initrd (cpio style, not using squashfs). So one option would be to silence **switch_root** in the **init** hook on livecd-kernel-image. Archlinux don't has it as they use **dmsnapshot**.
philm commented 2015-09-20 00:36:43 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: NuLogicSystems

Excuse me for not being knowledgeable on such things, but what is the advantage of squashfs?
And is there another option other than switch_root?

*Created by: NuLogicSystems* Excuse me for not being knowledgeable on such things, but what is the advantage of squashfs? And is there another option other than switch_root?
philm commented 2015-09-20 00:48:11 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: philmmanjaro

switch_root root works, but creates those warnings with squashfs. However using dmsnapshot in combination seems to make it work. However aufs and overlayfs might have some issues as you can read ;)

*Created by: philmmanjaro* **switch_root** root works, but creates those warnings with [squashfs](http://www.tldp.org/HOWTO/SquashFS-HOWTO/whatis.html). However using **dmsnapshot** in combination seems to make it work. However **aufs** and **overlayfs** might have some issues as you can read ;)
philm commented 2015-09-20 00:50:04 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: NuLogicSystems

So, have you decided how to proceed on this issue yet?

*Created by: NuLogicSystems* So, have you decided how to proceed on this issue yet?
philm commented 2015-09-20 00:51:34 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: philmmanjaro

There is no issue. It is just a warning message ...

*Created by: philmmanjaro* There is no issue. It is just a warning message ...
philm commented 2015-09-20 00:52:01 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: NuLogicSystems

Yes, but is there a way we can suppress it for now?

*Created by: NuLogicSystems* Yes, but is there a way we can suppress it for now?
philm commented 2015-09-20 00:52:41 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: philmmanjaro

What do you think I'm trying to do for 3 hours?

*Created by: philmmanjaro* What do you think I'm trying to do for 3 hours?
philm commented 2015-09-20 00:54:05 +00:00 (Migrated from gitlab2.manjaro.org)

Created by: NuLogicSystems

I wasn't saying that, I know you've been working hard on it.
I was just wondering if you've found anything yet. ;)
I do apologize if I unintentionally offended.

*Created by: NuLogicSystems* I wasn't saying that, I know you've been working hard on it. I was just wondering if you've found anything yet. ;) I do apologize if I unintentionally offended.
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
tools/manjaro-tools#137
No description provided.