On 18/7/19 8:49 pm, Doug Laidlaw wrote:

OK, but I have forgotten how to disable ipv6.  Pages on the Web for Fedora-style distros use a directory

/proc/sys/net/ipv6/conf/

That folder doesn't exist on my system. There is only one for ipv4.

But it seems that I do have ipv6 comnnectivity:

$ip a

<snip>

link/ether 40:8d:5c:47:38:79 brd ff:ff:ff:ff:ff:ff

</snip>

The good Doktor says:


"Re: Disable IPv6

Unread postby doktor5000 » Jul 27th, '15, 03:03
Use draknetcenter => advanced settings => disable IPv6. http://doc.mageia.org/mcc/5/en/content/ ... ml#d4e1606"

But that has changed as well. These look hopeful:

/etc/sysconfig/network-scripts/init.ipv6-global /etc/sysconfig/network-scripts/network-functions /etc/sysconfig/network-scripts/network-functions-ipv6

Doug.

--- MBSE BBS v1.0.7.12 (GNU/Linux-x86_64)
* Origin: Aioe.org NNTP Server (2:250/1@fidonet)

On Sun, 21 Jul 2019 00:09:25 +1000, Doug Laidlaw wrote:

On 18/7/19 8:49 pm, Doug Laidlaw wrote:

OK, but I have forgotten how to disable ipv6.  Pages on the Web for
Fedora-style distros use a directory

/proc/sys/net/ipv6/conf/

That folder doesn't exist on my system. There is only one for ipv4.

But it seems that I do have ipv6 comnnectivity:

$ip a

<snip>

link/ether 40:8d:5c:47:38:79 brd ff:ff:ff:ff:ff:ff

</snip>

The good Doktor says:

"Re: Disable IPv6

Unread postby doktor5000 » Jul 27th, '15, 03:03
Use draknetcenter => advanced settings => disable IPv6. http://doc.mageia.org/mcc/5/en/content/ ... ml#d4e1606"

But that has changed as well. These look hopeful:

/etc/sysconfig/network-scripts/init.ipv6-global /etc/sysconfig/network-scripts/network-functions /etc/sysconfig/network-scripts/network-functions-ipv6

Far be it above me to tangle with doktor5000 unless I absolutely knew
I am right on a topic.

After you make your changes, I would like to see the output from a
root terminal of the following:

lsof -i -n | grep -iE ipv6
journalctl --no-hostname | grep ipv6 | grep -v named


FYI: for any lurkers
If you add your id to the systemd-journal group, you would not
need to get into a root terminal to see the system journal contents.
You will need to log out/in to pick the group change.


--- MBSE BBS v1.0.7.12 (GNU/Linux-x86_64)
* Origin: A noiseless patient Spider (2:250/1@fidonet)

On Sat, 20 Jul 2019 12:16:15 -0400, Bit Twister <BitTwister@mouse-potato.com> wrote:

After you make your changes, I would like to see the output from a
root terminal of the following:

On my highly customized Mageia 6 x86_64 install, as I don't disable
ipv6 on my test installations ...

lsof -i -n | grep -iE ipv6

No output.

journalctl --no-hostname | grep ipv6 | grep -v named

# journalctl -b --no-hostname | grep ipv6 | grep -v named
Jul 19 16:05:13 kernel: Command line: BOOT_IMAGE=linux root=LABEL=x6b noiswmd modprobedebug rd.luks=0 rd.lvm=0 rd.md=0 rd.dm=0 ipv6.disable=1 audit=0 rd.driver.pre=ehci_hcd net.ifnames=0 vga=0x324
Jul 19 16:05:13 kernel: Kernel command line: BOOT_IMAGE=linux root=LABEL=x6b noiswmd modprobedebug rd.luks=0 rd.lvm=0 rd.md=0 rd.dm=0 ipv6.disable=1 audit=0 rd.driver.pre=ehci_hcd net.ifnames=0 vga=0x324

# grep IPV6 /etc/sysconfig/ifcfg-eth0
IPV6INIT=no
IPV6TO4INIT=no

Regards, Dave Hodgins

--
Change dwhodgins@nomail.afraid.org to davidwhodgins@teksavvy.com for
email replies.

--- MBSE BBS v1.0.7.12 (GNU/Linux-x86_64)
* Origin: A noiseless patient Spider (2:250/1@fidonet)

On Sat, 20 Jul 2019 14:08:21 -0400, David W. Hodgins wrote:

On Sat, 20 Jul 2019 12:16:15 -0400, Bit Twister

<BitTwister@mouse-potato.com> wrote:

After you make your changes, I would like to see the output from a
root terminal of the following:

On my highly customized Mageia 6 x86_64 install, as I don't disable
ipv6 on my test installations ...

lsof -i -n | grep -iE ipv6

No output.

journalctl --no-hostname | grep ipv6 | grep -v named

# journalctl -b --no-hostname | grep ipv6 | grep -v named
Jul 19 16:05:13 kernel: Command line: BOOT_IMAGE=linux root=LABEL=x6b

noiswmd modprobedebug rd.luks=0 rd.lvm=0 rd.md=0 rd.dm=0 ipv6.disable=1 audit=0 rd.driver.pre=ehci_hcd net.ifnames=0 vga=0x324

Jul 19 16:05:13 kernel: Kernel command line: BOOT_IMAGE=linux root=LABEL=x6b

noiswmd modprobedebug rd.luks=0 rd.lvm=0 rd.md=0 rd.dm=0 ipv6.disable=1 audit=0 rd.driver.pre=ehci_hcd net.ifnames=0 vga=0x324

Yep, yep, yep, I do the same kernel command line via grub "ipv6.disable=1" trick.

# grep IPV6 /etc/sysconfig/ifcfg-eth0
IPV6INIT=no
IPV6TO4INIT=no

That is odd, I had to use
grep IPV6 /etc/sysconfig/network-scripts/ifcfg-enp* :-)

What I was after was the results/work from using doktor5000 methodology.


--- MBSE BBS v1.0.7.12 (GNU/Linux-x86_64)
* Origin: A noiseless patient Spider (2:250/1@fidonet)

On 21/7/19 5:26 am, Bit Twister wrote:

What I was after was the results/work from using doktor5000 methodology.

Since doktor wrote, there have been changes to the relevant scripts.
"Disable ipv6" is no longer in the menu under "advanced settings." The
menu in 7 has hardly anything there.

My thought was to delay the startup of cups until everything else has
settled down. It would be extremely rare for me to try to print
something as soon as I log on. At the very least, I would want to write something to print.

--- MBSE BBS v1.0.7.12 (GNU/Linux-x86_64)
* Origin: Aioe.org NNTP Server (2:250/1@fidonet)

On Sun, 21 Jul 2019 05:52:49 +1000, Doug Laidlaw wrote:

On 21/7/19 5:26 am, Bit Twister wrote:

What I was after was the results/work from using doktor5000 methodology.

Since doktor wrote, there have been changes to the relevant scripts.
"Disable ipv6" is no longer in the menu under "advanced settings." The
menu in 7 has hardly anything there.

Just looked, saw the ipv4/6 tunnel was disabled.

My thought was to delay the startup of cups until everything else has
settled down. It would be extremely rare for me to try to print
something as soon as I log on. At the very least, I would want to write something to print.

Well, you are the sys admin.

Since my ISP does not give me an ipv6 address it is a waste of cpu resources
to have a ipv6/ipv4 stack/tunnel running.

Seeing articles of cracked web sites web page creating an ipv6 tunnel
back to criminal's site, makes it a no brainier to disable it.

Did you click up a root terminal to see what processes have an open ipv6 socket?
lsof -i -n | grep -iE ipv6

--- MBSE BBS v1.0.7.12 (GNU/Linux-x86_64)
* Origin: A noiseless patient Spider (2:250/1@fidonet)

On 21/7/19 6:58 am, Bit Twister wrote:

Did you click up a root terminal to see what processes have an open ipv6 socket?
lsof -i -n | grep -iE ipv6

Using sudo, no output. Are we sure that ipv6 is the culprit?

There seem to be various ways of delaying a systemd unit. All have
caveats that I don't understand. The simplest method is to use
ExecStartPre:

[Service]
ExecStartPre=/bin/sleep 20

That worked. It is too soon to see any undesirable side effects.


--- MBSE BBS v1.0.7.12 (GNU/Linux-x86_64)
* Origin: Aioe.org NNTP Server (2:250/1@fidonet)

On Sat, 20 Jul 2019 15:26:17 -0400, Bit Twister <BitTwister@mouse-potato.com> wrote:

On Sat, 20 Jul 2019 14:08:21 -0400, David W. Hodgins wrote:

# journalctl -b --no-hostname | grep ipv6 | grep -v named
Jul 19 16:05:13 kernel: Command line: BOOT_IMAGE=linux root=LABEL=x6b noiswmd modprobedebug rd.luks=0 rd.lvm=0 rd.md=0 rd.dm=0 ipv6.disable=1 audit=0 rd.driver.pre=ehci_hcd net.ifnames=0 vga=0x324
# grep IPV6 /etc/sysconfig/ifcfg-eth0
IPV6INIT=no
IPV6TO4INIT=no

That is odd, I had to use
grep IPV6 /etc/sysconfig/network-scripts/ifcfg-enp* :-)

If you're referring to the eth0 instead of an enp* network interface name, that's due to the kernel command line option net.ifnames=0.

Since I only have one network interface on this system, I don't have to
worry about the interface detection order, so don't need systemd to rename
the interface based on it's location in the pci bus.

Again, this is only on the highly customized install that I normally use.
On all of my test installs I let it rename the interface, so on this
system it uses enp7s0. While it doesn't make any real difference what
the nic is called, I prefer to use eth0 when there's only one nic.

Regards, Dave Hodgins

--
Change dwhodgins@nomail.afraid.org to davidwhodgins@teksavvy.com for
email replies.

--- MBSE BBS v1.0.7.12 (GNU/Linux-x86_64)
* Origin: A noiseless patient Spider (2:250/1@fidonet)

On Sun, 21 Jul 2019 07:29:14 +1000, Doug Laidlaw wrote:

On 21/7/19 6:58 am, Bit Twister wrote:

Did you click up a root terminal to see what processes have an open ipv6
socket?
lsof -i -n | grep -iE ipv6

Using sudo, no output.

That indicates nothing is listening on a ipv6 port/socket.

Are we sure that ipv6 is the culprit?

We/I are just collecting information and where possible, turn off/disable unneeded items.

Curious was cups up and running prior to the lsof

There seem to be various ways of delaying a systemd unit. All have
caveats that I don't understand.

hehe, I hear that. I play with the ones that have the most verbiage
about what I want done. If works, make a change script. and forget
about it to free up brain cells for better use. :)

The simplest method is to use
ExecStartPre:
Example:

You might always keep in mind about lurkers and someone reading the
thread sometime in the future trying to solve the same problem.

[Service]
ExecStartPre=/bin/sleep 20

That worked. It is too soon to see any undesirable side effects.

I do sincerely hope you did not modify /usr/lib/systemd/system/cups.service Next systemd update might wipe out your changes. :(

systemd documentation indicates you should create Drop In files
whenever possible and not modify the /usr/lib/systemd/system/ files.

Simplest method to create a service Drop In file is run
systemctl edit cups
and insert
[Service]
ExecStartPre=/bin/sleep 20

save/exit, run
systemctl restart cups
systemctl status cups
and it should show the Drop In: file is being used.



--- MBSE BBS v1.0.7.12 (GNU/Linux-x86_64)
* Origin: A noiseless patient Spider (2:250/1@fidonet)

On 18/7/19 10:26 am, TJ wrote:

On 7/17/19 8:42 AM, Doug Laidlaw wrote:

On 17/7/19 10:21 pm, TJ wrote:

On 7/14/19 4:38 AM, Doug Laidlaw wrote:

Running Mageia 7 Official with Xfce4. Cups refuses to start.  I have >>>> ctried;

(a) MCC says it is set to boot.  I set it not to, then set it again
to boot.

(b) Putting the link to the service file in graphical.target.wants.

Neither one helped.  I have to start the service manually for each
session.

Sorry I didn't respond earlier. My Internet service has been nearly
worthless for over a month. It happens to be working today - for now.

I ran into this back in March with the M7 Beta 2, but it had been
reported back in January. See
https://bugs.mageia.org/show_bug.cgi?id=24189

It seems to be a "race" condition during the boot process, and is
hardware dependent. It went away for me after a major upgrade of the
affected hardware. The upgrade had been a coincidence. It had been
planned for a while before I saw the problem.

I passed the motherboard on to my brother when M7 went Official, but
he doesn't have a printer, so has not been affected.

TJ

Interesting.  Yes, I noticed the race mentioned in the printout, but
there have been no recent changes to my hardware.  I think that the
solution to a race (this came up a while ago on Bugzilla) is to alter
the timeouts in the .service file, but I will need to look up the
details.

I did not mean that the race started with the new hardware. On the
contrary, it went away.

As I stated in the bug, I had done quite a bit of printing with the
older hardware with whatever M7 was current at the time in February. The problem suddenly appeared in March after some extensive Cauldron
updates. Too many for my feeble ability to isolate the problem. And then
it went away for me, making it nearly impossible for me to pursue it.

TJ

My problem that started this thread, solved itself. Cups was starting
in Cauldron, but not in Official. That seemed very strange. Finally,
Cups refused to start at all. I uninstalled the Cups packages, then reinstalled them, and the problem went away.

Doug.

--- MBSE BBS v1.0.7.12 (GNU/Linux-x86_64)
* Origin: Aioe.org NNTP Server (2:250/1@fidonet)

On Mon, 19 Aug 2019 13:51:12 +1000, Doug Laidlaw wrote:

I uninstalled the Cups packages

How messy is that, Doug?

(I looked in MCC/Software just now, and a whole slew of installed
packages appeared after entering just "cups" in the Find window.)

Regards,

--
/\/\aurice
(Replace "nomail.afraid" by "bcs" to reply by email)

--- MBSE BBS v1.0.7.12 (GNU/Linux-x86_64)
* Origin: A noiseless patient Spider (2:250/1@fidonet)

On 20/8/19 12:27 am, Maurice wrote:

On Mon, 19 Aug 2019 13:51:12 +1000, Doug Laidlaw wrote:

I uninstalled the Cups packages

How messy is that, Doug?

(I looked in MCC/Software just now, and a whole slew of installed
packages appeared after entering just "cups" in the Find window.)

Regards,

Some are not dfirectly dependent, but there were quite a few. I
wouldn't have done it if my intention had not been to reinstall immediately.

I filed Bug 25406. That has been marked as a duplicate of Bug 24189,
which was filed against Mga 7 while 7 was only Cauldron.

--
The complete truth is not the prerogative of the human judge.
- Judge Meir Shamgar.

--- MBSE BBS v1.0.7.12 (GNU/Linux-x86_64)
* Origin: Aioe.org NNTP Server (2:250/1@fidonet)