systemd

[Auke Kok]

On 06/14/2011 04:51 AM, Dennis Veatch wrote:
> On Monday, June 13, 2011 6:29:55 PM Auke Kok wrote:
>> On 06/13/2011 08:08 AM, Dennis Veatch wrote:
>>> Auke,
>>>
>>> tried your recent systemd changes with the following;
>>>
>>> use sysvinit - no
>>> with gtk+-2 - no (caused build to fail)
>>
>> hmm, strange, I'll look into that.
>>
>>> and a partitioning scheme of / and swap.
>>>
>>> Nothing in /var/log/messages jumped out and wireless still worked.
>>>
>>> I did notice a lot of this, not sure what to make of them;
>>>
>>> cc1: warning: /include: No such file or directory
>>>
>>>     CC     src/libsystemd_core_la-service.lo
>>>
>>> cc1: warning: /include: No such file or directory
>>>
>>>     CC     src/libsystemd_core_la-automount.lo
>>>
>>> cc1: warning: /include: No such file or directory
>>>
>>>     CC     src/libsystemd_core_la-mount.lo
>>>
>>> cc1: warning: /include: No such file or directory
>>>
>>>     CC     src/libsystemd_core_la-swap.lo
>>>
>>> cc1: warning: /include: No such file or directory
>>>
>>>     CC     src/libsystemd_core_la-device.lo
>>>
>>> cc1: warning: /include: No such file or directory
>>>
>>>     CC     src/libsystemd_core_la-target.lo
>>>
>>> snip
>>>
>>> cc1: warning: /include: No such file or directory
>>>
>>>     CC     src/libsystemd_core_la-dbus-target.lo
>>>
>>> cc1: warning: /include: No such file or directory
>>>
>>>     CC     src/libsystemd_core_la-dbus-mount.lo
>>>
>>> cc1: warning: /include: No such file or directory
>>>
>>>     CC     src/libsystemd_core_la-dbus-automount.lo
>>>
>>> cc1: warning: /include: No such file or directory
>>>
>>>     CC     src/libsystemd_core_la-dbus-swap.lo
>>>
>>> Apparently none fatal as the box still boots.
>>>
>>> The last time I diddled with this on a multi-partition/multi-drive box
>>> things did not go well. I may try that again unless there are still some
>>> gotcha's even with your changes.
>>
>> I saw those too, it's from us somewhere passing an unusual prefix. Since
>> the compile succeds, no worries.
>>
>
> Not sure about that. Tried it from cli using several different incantations of
> --includedir and --oldincludedir such as; /, /usr, /usr/include and that
> warning persisted. I think it has something to do with their configure script.
>
>> The systemd install is getting a bit better. With sysv compat mode now
>> being able to be switched off, you'll actually start using most of
>> systemd's internal startup code (which does a ton and really efficient).
>>
>> Here are the few things left to do:
>>
>> - need to package replacement links to /sbin/init, initctl, reboot,
>> shutdown etc etc. in some package that conflicts with sysvinit
>>
>> - need to make sysvinit removable entirely (no use keeping it around)
>>
>> - add more service unit files (I did a few, needs more).
>>
>> User steps to perform when converting:
>>
>> - remove all VFS's from /etc/fstab. You should only have physical device
>> nodes here
>
> So for a fstab like this;
>
> # /etc/fstab: static file system information.
> #<file system>  <mount point>  <type>  <options>  <dump>  <pass>
>
> # proc is required pretty much on any linux system except embedded systems:
> proc    /proc           proc    defaults        0 0
>
> # shmfs is surpassed and obsolete
> #shm    /dev/shm        shm     defaults        0 0
> none    /dev/shm        tmpfs   defaults        0 0
>
> # devfs and devpts are not friends, yet 2.6 needs devpts:
> devpts  /dev/pts        devpts  defaults        0 0
>
> # usb devfs can be helpfull on interactive (desktop) machines
> #usbfs  /proc/bus/usb   usbfs defaults       0 0
>
> # /tmp should be wiped on boot so by default it helps to have it on tmpfs
> tmpfs   /tmp            tmpfs   defaults,size=2048m,nr_inodes=64m       0 0
>
> # /var/lock and /var/run need to be clean on reboot:
> tmpfs   /var/lock       tmpfs   size=4m 0 0
> tmpfs   /var/run        tmpfs   size=4m 0 0
>
> # /var/tmp should NOT be on tmpfs, LSB states that the contents of it must
> # stay intact on reboot
> #tmpfs  /var/tmp        tmpfs   size=32m        0 0
>
>
>
> UUID=a26179bd-93de-4f4a-bffc-d22fab04614c       /       ext4    defaults
> 0 0
> UUID=a883786d-56c2-48a8-af33-31d3c1e88761       swap    swap    defaults
> 0 0
>
> Get rid of everything except for the last two lines?

yes

>> - ln -sf /proc/self/mounts /etc/mtab
>
> Not sure why that should or needs to be left for the user to correct.

because the lunar mount script will undo this every time it boots. (you 
can't fix it automatically).

Auke