[OT] Systemd fun fun fun

[Yrrah]

(x-posted: alt.comp.freeware,alt.os.linux.debian,alt.os.linux)

http://www.openwall.com/lists/oss-security/2016/09/29/26
Anyone for a tweet ?

Follow up to: alt.os.linux.debian,alt.os.linux

Yrrah

[J.O. Aho]

http://www.openwall.com/lists/oss-security/2016/09/29/26
Anyone for a tweet ?

It's kind of expected when you have an application which assumes
everything went well when it failed to get a result.

At least I sleep better knowing my more important systems do not use
systemd.

[Mr. Man-wai Chang]

Post by J.O. Aho
At least I sleep better knowing my more important systems do not use
systemd.

Well, a messy but compulsory systemd can create jobs. :)

[J.O. Aho]

Post by Mr. Man-wai Chang

Post by J.O. Aho
At least I sleep better knowing my more important systems do not use
systemd.

Well, a messy but compulsory systemd can create jobs. :)

Sure it could, but that's like working for microsoft support :)

[Mr. Man-wai Chang]

Post by J.O. Aho
Sure it could, but that's like working for microsoft support :)

Not limited to Micro$oft... it's called obfuscation! ;)

[Yrrah]

(x-posted: alt.comp.freeware,alt.os.linux.debian,alt.os.linux)

Post by Yrrah

http://www.openwall.com/lists/oss-security/2016/09/29/26
Anyone for a tweet ?

Follow up to: alt.os.linux.debian,alt.os.linux

"Most distributions running systemd have published patches and the
issue has been fixed in the upstream systemd code. While the bug was
quickly fixed, its existence fuelled the fires of the systemd
controversy on many message boards and raised concerns about systemd's
complexity."
https://distrowatch.com/weekly.php?issue=20161003#news

Yrrah

[Ivan Shmakov]

Post by Yrrah

http://www.openwall.com/lists/oss-security/2016/09/29/26
Anyone for a tweet ?

"Most distributions running systemd have published patches and the
issue has been fixed in the upstream systemd code. While the bug was
quickly fixed, its existence fuelled the fires of the systemd
controversy on many message boards and raised concerns about
systemd's complexity."
https://distrowatch.com/weekly.php?issue=20161003#news

JFTR, the best reading material I've found so far on Systemd are
these two blog posts dating back to 2014:

Systemd has 6 service startup notification types, and they're all wrong
http://ewontfix.com/15

Broken by design: systemd
http://ewontfix.com/14

I think these provide nice summaries on what to expect of this
specific part of the "modern" GNU/Linux system design.

[Yrrah]

Post by Ivan Shmakov
JFTR, the best reading material I've found so far on Systemd are
Systemd has 6 service startup notification types, and they're all wrong
Broken by design: systemd

Let me rephrase: these two blogs confirm your thoughts on the subject.

Yrrah

[Ivan Shmakov]

Post by Yrrah

Post by Ivan Shmakov
JFTR, the best reading material I've found so far on Systemd are
Systemd has 6 service startup notification types, and they're all wrong
Broken by design: systemd

Let me rephrase: these two blogs confirm your thoughts on the
subject.

I do not care much about Systemd, and as such, my thoughts on it
are few and far apart. The posts confirmed some of the things
I've suspected back in the day, and otherwise appear to use
argumentation I find believable.

I'm not going to pretend I can read others' minds and say that
these posts will make sense to anyone else. But I find it a
possibility still.

[Ivan Shmakov]

[...]

I'm not going to pretend I can read others' minds and say that these
posts will make sense to anyone else. But I find it a possibility
still.

With that disclaimer in mind, my list of favorites regarding
Systemd just got another entry -- this time (a what I believe
is) a nice summary of the reasons the "Init controversy" isn't
going to go away anytime soon.

"[Why] it's quite difficult to do a good job at supporting multiple
init systems" [quote; no discernible title originally; --IS.]
by Russ Allbery
https://lists.debian.org/msgid-search/***@hope.eyrie.org news:sbMAF-2mW-***@gated-at.bofh.it (Usenet) news:***@hope.eyrie.org (Gmane)

[Innocuous]

Post by Ivan Shmakov
Broken by design: systemd
http://ewontfix.com/14
I think these provide nice summaries on what to expect of this
specific part of the "modern" GNU/Linux system design.

I do not like systemd. I've changed over to using Devuan.
The concept of systemd is wrong in my view.

[William Unruh]

Post by Innocuous

Post by Ivan Shmakov
Broken by design: systemd
http://ewontfix.com/14
I think these provide nice summaries on what to expect of this
specific part of the "modern" GNU/Linux system design.

I do not like systemd. I've changed over to using Devuan.
The concept of systemd is wrong in my view.

Not sure the concept is wrong. Fast booting is a good idea. The implimentation has problems.

[J.O. Aho]

Post by William Unruh

Post by Innocuous

Post by Ivan Shmakov
Broken by design: systemd
http://ewontfix.com/14
I think these provide nice summaries on what to expect of this
specific part of the "modern" GNU/Linux system design.

I do not like systemd. I've changed over to using Devuan.
The concept of systemd is wrong in my view.

Not sure the concept is wrong. Fast booting is a good idea. The implimentation has problems.

There are other init systems which supports parallel starting of
services without the faults. systemd do not start everything, some
things it just leaves for later, so you may have a flawed system.

[Joe]

On Sat, 28 Jan 2017 18:55:04 -0000 (UTC)

Post by William Unruh

Post by Innocuous

Post by Ivan Shmakov
Broken by design: systemd
http://ewontfix.com/14
I think these provide nice summaries on what to expect of
this specific part of the "modern" GNU/Linux system design.

I do not like systemd. I've changed over to using Devuan.
The concept of systemd is wrong in my view.

Not sure the concept is wrong. Fast booting is a good idea. The
implimentation has problems.

As was said, parallel booting appeared some time before systemd, in
sid, at least.

Fast booting would be nice, but I'll believe it when I see it. Some
months ago, to add insult to injury, grub started adding about a
ten-second timeout ('Press any key..') before loading the initrd, and I
still haven't found how to stop that, other than hovering over it
waiting to hit a key. It ought to be something simple....

[Richard Kettlewell]

Post by Joe

Post by William Unruh
Not sure the concept is wrong. Fast booting is a good idea. The
implimentation has problems.

As was said, parallel booting appeared some time before systemd, in
sid, at least.

Dependency-based booting (which is the mechanism supporting parallel
booting) was introduced in Debian 6.0 (squeeze), released in 2011. It
wasn’t just about performance or parallelism, the old system based on
statically configured sequence numbering was hopelessly fragile.

Post by Joe
Fast booting would be nice, but I'll believe it when I see it. Some
months ago, to add insult to injury, grub started adding about a
ten-second timeout ('Press any key..') before loading the initrd, and I
still haven't found how to stop that, other than hovering over it
waiting to hit a key. It ought to be something simple....

https://www.gnu.org/software/grub/manual/html_node/Simple-configuration.html

[Carlos E.R.]

Post by Joe
Fast booting would be nice, but I'll believe it when I see it. Some
months ago, to add insult to injury, grub started adding about a
ten-second timeout ('Press any key..') before loading the initrd, and I
still haven't found how to stop that, other than hovering over it
waiting to hit a key. It ought to be something simple....

In openSUSE - check if it is the same in your distro:

/etc/default/grub:

GRUB_TIMEOUT=8

Then run "grub2-mkconfig -o /boot/grub2/grub.cfg" to apply.

[Joe]

On Sun, 29 Jan 2017 02:43:05 +0100

Post by Carlos E.R.

Post by Joe
Fast booting would be nice, but I'll believe it when I see it. Some
months ago, to add insult to injury, grub started adding about a
ten-second timeout ('Press any key..') before loading the initrd,
and I still haven't found how to stop that, other than hovering
over it waiting to hit a key. It ought to be something simple....

GRUB_TIMEOUT=8
Then run "grub2-mkconfig -o /boot/grub2/grub.cfg" to apply.

My grub timeout is on 5 and works as expected.

In recent months, the *next* screen says 'loading ramdisc' or similar,
followed by 'press any key' and sits there for ten seconds or so if I
don't press a key. I haven't found anything anywhere which seems to
refer to this.

[J.O. Aho]

Post by Joe
On Sun, 29 Jan 2017 02:43:05 +0100

Post by Carlos E.R.

Post by Joe
Fast booting would be nice, but I'll believe it when I see it. Some
months ago, to add insult to injury, grub started adding about a
ten-second timeout ('Press any key..') before loading the initrd,
and I still haven't found how to stop that, other than hovering
over it waiting to hit a key. It ought to be something simple....

GRUB_TIMEOUT=8
Then run "grub2-mkconfig -o /boot/grub2/grub.cfg" to apply.

My grub timeout is on 5 and works as expected.
In recent months, the *next* screen says 'loading ramdisc' or similar,
followed by 'press any key' and sits there for ten seconds or so if I
don't press a key. I haven't found anything anywhere which seems to
refer to this.

This would be in dracut settings, which is the one used in many
distribution to build the initramfs image. Most likely a change in a
script your distro uses or you have by mistake changed.

[Joe]

On Sun, 29 Jan 2017 12:40:49 +0100

Post by J.O. Aho

Post by Joe
On Sun, 29 Jan 2017 02:43:05 +0100

Post by Carlos E.R.

Post by Joe
Fast booting would be nice, but I'll believe it when I see it.
Some months ago, to add insult to injury, grub started adding
about a ten-second timeout ('Press any key..') before loading the
initrd, and I still haven't found how to stop that, other than
hovering over it waiting to hit a key. It ought to be something
simple....

GRUB_TIMEOUT=8
Then run "grub2-mkconfig -o /boot/grub2/grub.cfg" to apply.

My grub timeout is on 5 and works as expected.
In recent months, the *next* screen says 'loading ramdisc' or
similar, followed by 'press any key' and sits there for ten seconds
or so if I don't press a key. I haven't found anything anywhere
which seems to refer to this.

This would be in dracut settings, which is the one used in many
distribution to build the initramfs image. Most likely a change in a
script your distro uses or you have by mistake changed.

Thatnks for the try, but dracut isn't installed here and
initramfs-tools is used to create the initrd. I can find nothing
in /etc/initramfs-tools referring to a time.

There's no /etc/default entry for initramfs-tools, and the grub entry
contains only its own timeout.

[J.O. Aho]

Post by Joe
On Sun, 29 Jan 2017 12:40:49 +0100

Post by J.O. Aho

Post by Joe
On Sun, 29 Jan 2017 02:43:05 +0100

Post by Carlos E.R.

Post by Joe
Fast booting would be nice, but I'll believe it when I see it.
Some months ago, to add insult to injury, grub started adding
about a ten-second timeout ('Press any key..') before loading the
initrd, and I still haven't found how to stop that, other than
hovering over it waiting to hit a key. It ought to be something
simple....

GRUB_TIMEOUT=8
Then run "grub2-mkconfig -o /boot/grub2/grub.cfg" to apply.

My grub timeout is on 5 and works as expected.
In recent months, the *next* screen says 'loading ramdisc' or
similar, followed by 'press any key' and sits there for ten seconds
or so if I don't press a key. I haven't found anything anywhere
which seems to refer to this.

This would be in dracut settings, which is the one used in many
distribution to build the initramfs image. Most likely a change in a
script your distro uses or you have by mistake changed.

Thatnks for the try, but dracut isn't installed here and
initramfs-tools is used to create the initrd.

Should be possible with that too.

Post by Joe
I can find nothing
in /etc/initramfs-tools referring to a time.

I would take a look at the script it may build into the initramfs, you
could take the one you have and unpack it and see what's inside, if you
don't seem to find it there then I'm not sure.

Post by Joe
There's no /etc/default entry for initramfs-tools, and the grub entry
contains only its own timeout.

When reread your original post, you said it was before initrd, which
then should rule out the initramfs file.

Have you looked at the scripts in /etc/grub.d ? those are used to make
the entries in the grub menu.

[Joe]

On Sun, 29 Jan 2017 20:50:02 +0100

Post by J.O. Aho

Post by Joe
On Sun, 29 Jan 2017 12:40:49 +0100

Post by J.O. Aho

Post by Joe
In recent months, the *next* screen says 'loading ramdisc' or
similar, followed by 'press any key' and sits there for ten
seconds or so if I don't press a key. I haven't found anything
anywhere which seems to refer to this.

This would be in dracut settings, which is the one used in many
distribution to build the initramfs image. Most likely a change in
a script your distro uses or you have by mistake changed.

Thatnks for the try, but dracut isn't installed here and
initramfs-tools is used to create the initrd.

Should be possible with that too.

Post by Joe
I can find nothing
in /etc/initramfs-tools referring to a time.

I would take a look at the script it may build into the initramfs, you
could take the one you have and unpack it and see what's inside, if
you don't seem to find it there then I'm not sure.

Post by Joe
There's no /etc/default entry for initramfs-tools, and the grub
entry contains only its own timeout.

When reread your original post, you said it was before initrd, which
then should rule out the initramfs file.
Have you looked at the scripts in /etc/grub.d ? those are used to make
the entries in the grub menu.

OK, a bit of light is shed. I'm seeing these lines after the grub menu
selection or timeout:

oading Linux <kernel number>...
oading initial ramdisk...

ress any key to continue...


I'm making the assumption that the monitor has not yet synchronised and
there are the letters L, L and P at the beginning of these three lines.

If I make the assumption that there is also a vertical error, leave the
system sitting on the grub menu until I manually pull the sync across
and down, then go onto the next page, I see the line:

no symbol table

above the others. Light dawns. I still don't know what the problem is,
but I now know it isn't some new configuration that I'm missing, it's
something broken. It looks like this:

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=841297

except I don't have grub-efi. Poking around further, some people say
it can be fixed by repairing or reinstalling grub, but it's not causing
me enough trouble to be worth the risk. Presumably when a new grub2
arrives, it will sort itself out.

Hope this helps someone else, at least about the monitor gotcha...

[Carlos E.R.]

Post by Joe
On Sun, 29 Jan 2017 02:43:05 +0100

Post by Carlos E.R.

Post by Joe
Fast booting would be nice, but I'll believe it when I see it. Some
months ago, to add insult to injury, grub started adding about a
ten-second timeout ('Press any key..') before loading the initrd,
and I still haven't found how to stop that, other than hovering
over it waiting to hit a key. It ought to be something simple....

GRUB_TIMEOUT=8
Then run "grub2-mkconfig -o /boot/grub2/grub.cfg" to apply.

My grub timeout is on 5 and works as expected.
In recent months, the *next* screen says 'loading ramdisc' or similar,
followed by 'press any key' and sits there for ten seconds or so if I
don't press a key. I haven't found anything anywhere which seems to
refer to this.

Then it seems like some problem it finds and waits for your intervention.

[Joe]

On Mon, 30 Jan 2017 04:23:31 +0100

Post by Carlos E.R.
Then it seems like some problem it finds and waits for your
intervention.

Yes, I now know it is a 'no symbol table' error, though there doesn't
seem to be any trouble.

[Carlos E.R.]

Post by Joe
On Mon, 30 Jan 2017 04:23:31 +0100

Post by Carlos E.R.
Then it seems like some problem it finds and waits for your
intervention.

Yes, I now know it is a 'no symbol table' error, though there doesn't
seem to be any trouble.

Google finds hits. I looked at one:

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1633839

seems to indicate a grub bug.

[Joe]

On Mon, 30 Jan 2017 14:12:25 +0100

Post by Carlos E.R.

Post by Joe
On Mon, 30 Jan 2017 04:23:31 +0100

Post by Carlos E.R.
Then it seems like some problem it finds and waits for your
intervention.

Yes, I now know it is a 'no symbol table' error, though there
doesn't seem to be any trouble.

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1633839
seems to indicate a grub bug.

Thanks, yes, I found that page and got the impression that some people
had solved it by reinstalling grub, but that had not worked for others.
I'm on LVM for all except /boot, which may be relevant or may not.

It's not really serious enough to mess around with, as everything still
seems to work OK. I have problems, but they are the same ones I had
before this started happening, which could well have been in October.
If it was causing other people problems, it would probably have been
fixed by now.

[Cybe R. Wizard]

On Mon, 30 Jan 2017 14:12:25 +0100

Post by Carlos E.R.

Post by Joe
On Mon, 30 Jan 2017 04:23:31 +0100

Post by Carlos E.R.
Then it seems like some problem it finds and waits for your
intervention.

Yes, I now know it is a 'no symbol table' error, though there
doesn't seem to be any trouble.

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1633839
seems to indicate a grub bug.

I thought I'd heard that expression before:
----------
<https://duckduckgo.com/?q=grub+bug&t=midori&ia=web>
...and images:
<https://duckduckgo.com/?q=grub+bug&t=midori&iax=1&ia=images>
----------

Big smile.

Cybe R. Wizard

[Shadow]

On Sun, 29 Jan 2017 02:43:05 +0100, "Carlos E.R."

Post by Carlos E.R.

Post by Joe
Fast booting would be nice, but I'll believe it when I see it. Some
months ago, to add insult to injury, grub started adding about a
ten-second timeout ('Press any key..') before loading the initrd, and I
still haven't found how to stop that, other than hovering over it
waiting to hit a key. It ought to be something simple....

GRUB_TIMEOUT=8
Then run "grub2-mkconfig -o /boot/grub2/grub.cfg" to apply.

Not just:
update-grub
???
How complicated.
[]'s

[Chris Ahlstrom]

Post by Shadow
On Sun, 29 Jan 2017 02:43:05 +0100, "Carlos E.R."

Post by Carlos E.R.

Post by Joe
Fast booting would be nice, but I'll believe it when I see it. Some
months ago, to add insult to injury, grub started adding about a
ten-second timeout ('Press any key..') before loading the initrd, and I
still haven't found how to stop that, other than hovering over it
waiting to hit a key. It ought to be something simple....

GRUB_TIMEOUT=8
Then run "grub2-mkconfig -o /boot/grub2/grub.cfg" to apply.

update-grub
???
How complicated.
[]'s

It's pretty simple, and not related to systemd. (I use roughly the same
on CentOS.)

[William Unruh]

Post by Joe
On Sat, 28 Jan 2017 18:55:04 -0000 (UTC)

Post by William Unruh

Post by Innocuous

Post by Ivan Shmakov
Broken by design: systemd
http://ewontfix.com/14
I think these provide nice summaries on what to expect of
this specific part of the "modern" GNU/Linux system design.

I do not like systemd. I've changed over to using Devuan.
The concept of systemd is wrong in my view.

Not sure the concept is wrong. Fast booting is a good idea. The
implimentation has problems.

As was said, parallel booting appeared some time before systemd, in
sid, at least.
Fast booting would be nice, but I'll believe it when I see it. Some
months ago, to add insult to injury, grub started adding about a
ten-second timeout ('Press any key..') before loading the initrd, and I
still haven't found how to stop that, other than hovering over it
waiting to hit a key. It ought to be something simple....

Nothing to do with systemd.
Edit the line in /boot/grub2/
set timeout=
to whatever you want. I have mine at 5 so I can if I wish boot windows.
But make it whatever you want.

[Joe]

On Sun, 29 Jan 2017 07:00:37 -0000 (UTC)

Post by William Unruh

Post by Joe
On Sat, 28 Jan 2017 18:55:04 -0000 (UTC)

Post by William Unruh

Post by Innocuous

Post by Ivan Shmakov
Broken by design: systemd
http://ewontfix.com/14
I think these provide nice summaries on what to expect
of this specific part of the "modern" GNU/Linux system design.

I do not like systemd. I've changed over to using Devuan.
The concept of systemd is wrong in my view.

Not sure the concept is wrong. Fast booting is a good idea. The
implimentation has problems.

As was said, parallel booting appeared some time before systemd, in
sid, at least.
Fast booting would be nice, but I'll believe it when I see it. Some
months ago, to add insult to injury, grub started adding about a
ten-second timeout ('Press any key..') before loading the initrd,
and I still haven't found how to stop that, other than hovering
over it waiting to hit a key. It ought to be something simple....

Nothing to do with systemd.

I wasn't suggesting that it was, just that systemd doesn't
automatically deliver a fast boot, and this phenomenon has slowed it
even more unless I give it a kick.

Post by William Unruh
Edit the line in /boot/grub2/
set timeout=
to whatever you want. I have mine at 5 so I can if I wish boot
windows. But make it whatever you want.

Possibly it is only in sid so far.

My grub timeout is on 5 and works as expected.

In recent months, the *next* screen says 'loading ramdisc' or similar,
followed by 'press any key' and sits there for ten seconds or so if I
don't press a key. I haven't found anything anywhere which seems to
refer to this.

[Whiskers]

Post by William Unruh

Broken by design: systemd http://ewontfix.com/14
I think these provide nice summaries on what to expect of this
specific part of the "modern" GNU/Linux system design.

I do not like systemd. I've changed over to using Devuan. The
concept of systemd is wrong in my view.

Not sure the concept is wrong. Fast booting is a good idea. The
implimentation has problems.

Systemd does seem to work pretty well. I had fun when Arch Linux
changed over to it and I had some re-configuring to do on my humble
laptop but it all went well. I'm sure it had nothing to do with the
hardware failure that happened much later; the machine was over ten
years old.

This new machine is the first one I've used that has UEFI (and 'secure
boot') to cope with, as well as being the first one I've installed Arch
Linux on with Systemd from the start. It isn't a dual boot, I gave
Windows the heave-ho, so that's one complication I could forget. I
chose to use systemd-boot (formerly gummiboot) on its own. Setting up
seemed to be a little more logically consistent than the previous init
system - but also rather more complex. Fortunately Arch Linux has
excellent documentation and guidance.

Booting is certainly quick. I do get an occasional problem such as the
X server or some part of the graphical stuff failing to run, but I don't
know if that's down to Systemd or to the 'hybrid hard disc' interfering
with the order in which things start.

I'm glad I don't have to run anything bigger than this laptop though.

[Chester A. Arthur]

Post by Innocuous
I do not like systemd. I've changed over to using Devuan.
The concept of systemd is wrong in my view.

Can anyone give a comparison between Devuan and MX-16 ?
Both are free of SysD. MX-16 says they have stubs to
handle the details. I hear Devuan is a BIG rewrite
especially of DBus, and has some problems because of it.

Thinking of installing MX-16 as a test.

[senior-student]

Post by Innocuous
I do not like systemd. I've changed over to using Devuan.
The concept of systemd is wrong in my view.

Can anyone give a comparison between Devuan and MX-16 ? Both are free of
SysD. MX-16 says they have stubs to handle the details. I hear Devuan is
a BIG rewrite especially of DBus, and has some problems because of it.
Thinking of installing MX-16 as a test.

try pclinuxos