[opensuse-factory] changing runlevel drops internet connection

newer
[opensuse-factory] suse-prime-alt...

Patrick Shanahan

10 Feb 2016 10 Feb '16

21:12

I have five local networked Tw machines which all drop network when the runlevel/multi-user/graphical target is changed. Plays havoc with remote access. One with a wired interface retains network but is inaccessable as external sites are firewalled requiring a SuSEfirewall2 restart. The wireless systems require require a delayed restart of wickedd and a reconnect. -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Show replies by date

Patrick Shanahan

10 Feb 10 Feb

21:18

* Patrick Shanahan [02-10-16 16:14]:

...

I have five local networked Tw machines which all drop network when the runlevel/multi-user/graphical target is changed. Plays havoc with remote access.

One with a wired interface retains network but is inaccessable as external sites are firewalled requiring a SuSEfirewall2 restart.

The wireless systems require require a delayed restart of wickedd and a reconnect.

changing runlevel on a wireless laptop also logs out of the tty -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Andrei Borzenkov

11 Feb 11 Feb

03:30

11.02.2016 00:18, Patrick Shanahan пишет:

...

* Patrick Shanahan [02-10-16 16:14]:

...
I have five local networked Tw machines which all drop network when the runlevel/multi-user/graphical target is changed. Plays havoc with remote access.

There is discussion about it in forums and I cannot reproduce it or you need to give more technical explanation what "drop network" is.

...

...
One with a wired interface retains network but is inaccessable as external sites are firewalled requiring a SuSEfirewall2 restart.

Which makes me suspect that you use NetworkManager with user connection on another system. Is it true?

...

...
The wireless systems require require a delayed restart of wickedd and a reconnect.

changing runlevel on a wireless laptop also logs out of the tty

I observe this as well. Technically this behavior matches documentation; I do not know if this is intentional though. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Patrick Shanahan

04:33

* Andrei Borzenkov [02-10-16 22:32]:

...

11.02.2016 00:18, Patrick Shanahan пишет:

...
* Patrick Shanahan [02-10-16 16:14]:

...
I have five local networked Tw machines which all drop network when the runlevel/multi-user/graphical target is changed. Plays havoc with remote access.

There is discussion about it in forums and I cannot reproduce it or you need to give more technical explanation what "drop network" is.

My Toshiba Satellite S55-C5274 laptop goes from having wireless network connection to having no network connection after changing from or to multi-user and graphical targets. *Usually* restarting wickedd will restore network connection. But I also have two other laptops and a desktop with wireless connection which exhibit the same condition.

...

...
...
One with a wired interface retains network but is inaccessable as external sites are firewalled requiring a SuSEfirewall2 restart.

Which makes me suspect that you use NetworkManager with user connection on another system. Is it true?

No, I *only* have wicked, NetworkManager only allows *me* connection with graphical target and I need network also using multi-user target.

...

...
...
The wireless systems require require a delayed restart of wickedd and a reconnect.

changing runlevel on a wireless laptop also logs out of the tty

I observe this as well. Technically this behavior matches documentation; I do not know if this is intentional though.

Being logged out of the tty has only been happening to me during the last week. Never that I can recall before that. These conditions cause me concern as I frequently require remote connection. I am willing to perform tests, one laptop is used primarily for testing anyway. tks, -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Andrei Borzenkov

05:25

Отправлено с iPhone

...

11 февр. 2016 г., в 7:33, Patrick Shanahan написал(а):

* Andrei Borzenkov [02-10-16 22:32]:

...
11.02.2016 00:18, Patrick Shanahan пишет:

...
* Patrick Shanahan [02-10-16 16:14]:

...
I have five local networked Tw machines which all drop network when the runlevel/multi-user/graphical target is changed. Plays havoc with remote access.

There is discussion about it in forums and I cannot reproduce it or you need to give more technical explanation what "drop network" is.

My Toshiba Satellite S55-C5274 laptop goes from having wireless network connection to having no network connection after changing from or to multi-user and graphical targets. *Usually* restarting wickedd will restore network connection.

I have a theory. Could you please get log with debug level and upload it somewhere (susepaste.org) Boot normally Switch on debug log level of systemd /bin/kill -RTMIN+22 1 (You need to use external kill command) Switch run level Upload output of "journalctl -b"

...

But I also have two other laptops and a desktop with wireless connection which exhibit the same condition.

...
...
...
One with a wired interface retains network but is inaccessable as external sites are firewalled requiring a SuSEfirewall2 restart.

Which makes me suspect that you use NetworkManager with user connection on another system. Is it true?

No, I *only* have wicked, NetworkManager only allows *me* connection with graphical target and I need network also using multi-user target.

...
...
...
The wireless systems require require a delayed restart of wickedd and a reconnect.

changing runlevel on a wireless laptop also logs out of the tty

I observe this as well. Technically this behavior matches documentation; I do not know if this is intentional though.

Being logged out of the tty has only been happening to me during the last week. Never that I can recall before that.

These conditions cause me concern as I frequently require remote connection.

I am willing to perform tests, one laptop is used primarily for testing anyway.

tks, -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Patrick Shanahan

13:45

* Andrei Borzenkov [02-11-16 00:28]:

...

Отправлено с iPhone

...
11 февр. 2016 г., в 7:33, Patrick Shanahan написал(а):

* Andrei Borzenkov [02-10-16 22:32]:

...
11.02.2016 00:18, Patrick Shanahan пишет:

...
* Patrick Shanahan [02-10-16 16:14]:

...
I have five local networked Tw machines which all drop network when the runlevel/multi-user/graphical target is changed. Plays havoc with remote access.

There is discussion about it in forums and I cannot reproduce it or you need to give more technical explanation what "drop network" is.

My Toshiba Satellite S55-C5274 laptop goes from having wireless network connection to having no network connection after changing from or to multi-user and graphical targets. *Usually* restarting wickedd will restore network connection.

I have a theory. Could you please get log with debug level and upload it somewhere (susepaste.org)

Boot normally Switch on debug log level of systemd

/bin/kill -RTMIN+22 1

(You need to use external kill command)

Switch run level Upload output of "journalctl -b"

see quite long journalctl.txt here: http://wahoo.no-ip.org/~pat/journalctl.txt tks, -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Andrei Borzenkov

17:39

11.02.2016 16:45, Patrick Shanahan пишет:

...

* Andrei Borzenkov [02-11-16 00:28]:

...
Отправлено с iPhone

...
11 февр. 2016 г., в 7:33, Patrick Shanahan написал(а):

* Andrei Borzenkov [02-10-16 22:32]:

...
11.02.2016 00:18, Patrick Shanahan пишет:

...
* Patrick Shanahan [02-10-16 16:14]:

...
I have five local networked Tw machines which all drop network when the runlevel/multi-user/graphical target is changed. Plays havoc with remote access.

There is discussion about it in forums and I cannot reproduce it or you need to give more technical explanation what "drop network" is.

My Toshiba Satellite S55-C5274 laptop goes from having wireless network connection to having no network connection after changing from or to multi-user and graphical targets. *Usually* restarting wickedd will restore network connection.

I have a theory. Could you please get log with debug level and upload it somewhere (susepaste.org)

Boot normally Switch on debug log level of systemd

/bin/kill -RTMIN+22 1

(You need to use external kill command)

Switch run level Upload output of "journalctl -b"

see quite long journalctl.txt here:

I asked you to provide log immediately after boot, not after running over three days. journalctl has --since option in case you did not notice yet. And I asked you to enable systemd debug log level that you did not do either, which makes this enormous log rather useless for bug report. But it still confirms what I suspected. If you now do systemctl enable wpa_supplicant.service does connection still drop? -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Patrick Shanahan

19:11

* Andrei Borzenkov [02-11-16 12:41]:

...

11.02.2016 16:45, Patrick Shanahan пишет:

...
* Andrei Borzenkov [02-11-16 00:28]:

...
Отправлено с iPhone

...
11 февр. 2016 г., в 7:33, Patrick Shanahan написал(а):

* Andrei Borzenkov [02-10-16 22:32]:

...
11.02.2016 00:18, Patrick Shanahan пишет:

...
* Patrick Shanahan [02-10-16 16:14]: > > I have five local networked Tw machines which all drop network when the > runlevel/multi-user/graphical target is changed. Plays havoc with remote > access.

There is discussion about it in forums and I cannot reproduce it or you need to give more technical explanation what "drop network" is.

My Toshiba Satellite S55-C5274 laptop goes from having wireless network connection to having no network connection after changing from or to multi-user and graphical targets. *Usually* restarting wickedd will restore network connection.

I have a theory. Could you please get log with debug level and upload it somewhere (susepaste.org)

Boot normally Switch on debug log level of systemd

/bin/kill -RTMIN+22 1

(You need to use external kill command)

Switch run level Upload output of "journalctl -b"

see quite long journalctl.txt here:

I asked you to provide log immediately after boot, not after running over three days. journalctl has --since option in case you did not notice yet.

And I asked you to enable systemd debug log level that you did not do either, which makes this enormous log rather useless for bug report.

But it still confirms what I suspected. If you now do

systemctl enable wpa_supplicant.service

does connection still drop?

No, wireless connection was maintained. systemctl enable wpa_supp.... revealed that there was no sym-link for multi-user.target. Do we need a bug report? ps: terminal where change from graphical to multi-user target was closed. Is this also a missing sym-link? tks, much -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Patrick Shanahan

19:22

* Patrick Shanahan [02-11-16 14:11]:

...

* Andrei Borzenkov [02-11-16 12:41]:

...
11.02.2016 16:45, Patrick Shanahan пишет:

...
* Andrei Borzenkov [02-11-16 00:28]:

...
Отправлено с iPhone

...
11 февр. 2016 г., в 7:33, Patrick Shanahan написал(а):

* Andrei Borzenkov [02-10-16 22:32]:

...
11.02.2016 00:18, Patrick Shanahan пишет: > * Patrick Shanahan [02-10-16 16:14]: >> >> I have five local networked Tw machines which all drop network when the >> runlevel/multi-user/graphical target is changed. Plays havoc with remote >> access.

There is discussion about it in forums and I cannot reproduce it or you need to give more technical explanation what "drop network" is.

My Toshiba Satellite S55-C5274 laptop goes from having wireless network connection to having no network connection after changing from or to multi-user and graphical targets. *Usually* restarting wickedd will restore network connection.

I have a theory. Could you please get log with debug level and upload it somewhere (susepaste.org)

Boot normally Switch on debug log level of systemd

/bin/kill -RTMIN+22 1

(You need to use external kill command)

Switch run level Upload output of "journalctl -b"

see quite long journalctl.txt here:

I asked you to provide log immediately after boot, not after running over three days. journalctl has --since option in case you did not notice yet.

And I asked you to enable systemd debug log level that you did not do either, which makes this enormous log rather useless for bug report.

But it still confirms what I suspected. If you now do

systemctl enable wpa_supplicant.service

does connection still drop?

No, wireless connection was maintained.

systemctl enable wpa_supp.... revealed that there was no sym-link for multi-user.target.

Do we need a bug report?

NOTE: appears to have solved dropping network on my other wireless connected local machines: systemctl enable wpa_supplicant Created symlink from /etc/systemd/system/multi-user.target.wants/wpa_supplicant.service to //usr/lib/systemd/system/wpa_supplicant.service.

...

ps: terminal where change from graphical to multi-user target was closed. Is this also a missing sym-link?

tks, much

-- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Andrei Borzenkov

12 Feb 12 Feb

03:33

11.02.2016 22:11, Patrick Shanahan пишет:

...

...
But it still confirms what I suspected. If you now do

systemctl enable wpa_supplicant.service

does connection still drop?

No, wireless connection was maintained.

OK, so here it was happens. By definition "systemctl isolate foo.target" stops all services that would not be pulled in by foo.target. Unfortunately this covers only static dependencies, that can be deduced from Wants/Requires configuration directives. But this does not take in account dynamic socket activated services. Now wpa_supplicant.service is disabled by default, but started on request when D-Bus connection is made. But that also means wpa_supplicant is considered as not belonging to any "run-level" and it is stopped. Feb 07 17:23:35 toshiba dbus[1074]: [system] Activating via systemd: service name='fi.epitest.hostap.WPASupplicant' unit='wpa_supplicant.service' Feb 07 17:23:35 toshiba dbus[1074]: [system] Successfully activated service 'fi.epitest.hostap.WPASupplicant' ... Feb 07 17:24:11 toshiba systemd[1]: Reached target Graphical Interface. ... Feb 09 08:07:22 toshiba systemd[1]: Stopping WPA Supplicant daemon... ... Feb 09 08:08:58 toshiba systemd[1]: Reached target Multi-User System. Stopping wpa_supplicant of course kills wireless connections. This really needs bug report. Brute force workaround is to enable wpa_supplicant service which will add dependency on multi-user.target so wpa_supplicant is no more killed. But it looks like fundamental design issue in systemd ...

...

systemctl enable wpa_supp.... revealed that there was no sym-link for multi-user.target.

Do we need a bug report?

By all means. But TBH I do not hold my breath - I do not see any way to mark such dynamic service as "in use" ...

...

ps: terminal where change from graphical to multi-user target was closed.

Yes, this is the same problem.

...

Is this also a missing sym-link?

Here it is worse. All user sessions run as part of so called "scope" units (look at systemd-cgls output). This scope units are created on the fly. It is simply impossible impossible to add any static dependency from multi-user.target to session-N.scope. And these scope units are apparently stopped as well when you do "systemctl isolate" causing all user processes in this session be stopped. So yes, but report is definitely needed. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Patrick Shanahan

19:07

* Andrei Borzenkov [02-11-16 22:35]:

...

11.02.2016 22:11, Patrick Shanahan пишет:

...
...
But it still confirms what I suspected. If you now do

systemctl enable wpa_supplicant.service

does connection still drop?

No, wireless connection was maintained.

OK, so here it was happens.

By definition "systemctl isolate foo.target" stops all services that would not be pulled in by foo.target. Unfortunately this covers only static dependencies, that can be deduced from Wants/Requires configuration directives. But this does not take in account dynamic socket activated services. Now wpa_supplicant.service is disabled by default, but started on request when D-Bus connection is made. But that also means wpa_supplicant is considered as not belonging to any "run-level" and it is stopped.

Feb 07 17:23:35 toshiba dbus[1074]: [system] Activating via systemd: service name='fi.epitest.hostap.WPASupplicant' unit='wpa_supplicant.service' Feb 07 17:23:35 toshiba dbus[1074]: [system] Successfully activated service 'fi.epitest.hostap.WPASupplicant' ... Feb 07 17:24:11 toshiba systemd[1]: Reached target Graphical Interface. ... Feb 09 08:07:22 toshiba systemd[1]: Stopping WPA Supplicant daemon... ... Feb 09 08:08:58 toshiba systemd[1]: Reached target Multi-User System.

Stopping wpa_supplicant of course kills wireless connections.

This really needs bug report. Brute force workaround is to enable wpa_supplicant service which will add dependency on multi-user.target so wpa_supplicant is no more killed. But it looks like fundamental design issue in systemd ...

...
systemctl enable wpa_supp.... revealed that there was no sym-link for multi-user.target.

Do we need a bug report?

By all means. But TBH I do not hold my breath - I do not see any way to mark such dynamic service as "in use" ...

...
ps: terminal where change from graphical to multi-user target was closed.

Yes, this is the same problem.

...
Is this also a missing sym-link?

Here it is worse. All user sessions run as part of so called "scope" units (look at systemd-cgls output). This scope units are created on the fly. It is simply impossible impossible to add any static dependency from multi-user.target to session-N.scope. And these scope units are apparently stopped as well when you do "systemctl isolate" causing all user processes in this session be stopped.

So yes, but report is definitely needed.

Bug 966535 https://bugzilla.opensuse.org/show_bug.cgi?id=966535 Referenced this thread in bug report. tks for your help -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Patrick Shanahan

18 Feb 18 Feb

13:23

New subject: [opensuse-factory] UPDATE (sorry situation) re: changing runlevel drops internet connection

* Patrick Shanahan [02-12-16 14:08]:

...

* Andrei Borzenkov [02-11-16 22:35]:

...
11.02.2016 22:11, Patrick Shanahan пишет:

...
...
But it still confirms what I suspected. If you now do

systemctl enable wpa_supplicant.service

does connection still drop?

No, wireless connection was maintained.

OK, so here it was happens.

By definition "systemctl isolate foo.target" stops all services that would not be pulled in by foo.target. Unfortunately this covers only static dependencies, that can be deduced from Wants/Requires configuration directives. But this does not take in account dynamic socket activated services. Now wpa_supplicant.service is disabled by default, but started on request when D-Bus connection is made. But that also means wpa_supplicant is considered as not belonging to any "run-level" and it is stopped.

Feb 07 17:23:35 toshiba dbus[1074]: [system] Activating via systemd: service name='fi.epitest.hostap.WPASupplicant' unit='wpa_supplicant.service' Feb 07 17:23:35 toshiba dbus[1074]: [system] Successfully activated service 'fi.epitest.hostap.WPASupplicant' ... Feb 07 17:24:11 toshiba systemd[1]: Reached target Graphical Interface. ... Feb 09 08:07:22 toshiba systemd[1]: Stopping WPA Supplicant daemon... ... Feb 09 08:08:58 toshiba systemd[1]: Reached target Multi-User System.

Stopping wpa_supplicant of course kills wireless connections.

This really needs bug report. Brute force workaround is to enable wpa_supplicant service which will add dependency on multi-user.target so wpa_supplicant is no more killed. But it looks like fundamental design issue in systemd ...

...
systemctl enable wpa_supp.... revealed that there was no sym-link for multi-user.target.

Do we need a bug report?

By all means. But TBH I do not hold my breath - I do not see any way to mark such dynamic service as "in use" ...

...
ps: terminal where change from graphical to multi-user target was closed.

Yes, this is the same problem.

...
Is this also a missing sym-link?

Here it is worse. All user sessions run as part of so called "scope" units (look at systemd-cgls output). This scope units are created on the fly. It is simply impossible impossible to add any static dependency from multi-user.target to session-N.scope. And these scope units are apparently stopped as well when you do "systemctl isolate" causing all user processes in this session be stopped.

So yes, but report is definitely needed.

Bug 966535 https://bugzilla.opensuse.org/show_bug.cgi?id=966535

Referenced this thread in bug report.

Some enlightening comments: https://bugzilla.opensuse.org/show_bug.cgi?id=966535#c3 (In reply to patrick shanahan from comment #2) Dr. Werner Fink 2016-02-15 13:26:21 UTC

...

Perhaps because openSUSE has multi-user.target and requires internet and there is no internet w/o wpa_supplicant. And it's runlevel 3.

Yep ... and both wpa_supplicant.service and wicked services are not part of systemd and therefore this bug does *not* belong to systemd. The systemd init simply does what the dependency rules do indicate. https://bugzilla.opensuse.org/show_bug.cgi?id=966535#c28 Dr. Werner Fink 2016-02-18 10:33:12 UTC https://github.com/systemd/systemd/issues/1969 Btw: do not use tumbleweed on production systems ;) https://bugzilla.opensuse.org/show_bug.cgi?id=966535#c29 Ludwig Nussel 2016-02-18 10:41:44 UTC (In reply to Dr. Werner Fink from comment #28)

...

https://github.com/systemd/systemd/issues/1969

Btw: do not use tumbleweed on production systems ;)

of course you should, otherwise we wouldn't find such bugs. The trick is to be smart and creative enough to workaround glitches :) -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Andrei Borzenkov

14:26

New subject: [opensuse-factory] UPDATE (sorry situation) re: changing runlevel drops internet connection

On Thu, Feb 18, 2016 at 4:23 PM, Patrick Shanahan wrote:

...

https://bugzilla.opensuse.org/show_bug.cgi?id=966535#c3

(In reply to patrick shanahan from comment #2) Dr. Werner Fink 2016-02-15 13:26:21 UTC

...
Perhaps because openSUSE has multi-user.target and requires internet and there is no internet w/o wpa_supplicant. And it's runlevel 3.

Yep ... and both wpa_supplicant.service and wicked services are not part of systemd and therefore this bug does *not* belong to systemd.

It does, although more upstream. Still, it is problem in systemd that affects multiple programs, so suggesting to workaround it in each individual program is not productive.

...

systemd init simply does what the dependency rules do indicate.

There is no way to express dynamic dependencies created by socket activation; and socket activation is marketed as one of systemd advantage. So it should be able to cope with this situation. Anyway, there is upstream issue opened as noted below. This affects more than one distribution.

...

https://bugzilla.opensuse.org/show_bug.cgi?id=966535#c28 Dr. Werner Fink 2016-02-18 10:33:12 UTC https://github.com/systemd/systemd/issues/1969

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Patrick Shanahan

15:30

New subject: [opensuse-factory] UPDATE (sorry situation) re: changing runlevel drops internet connection

* Andrei Borzenkov [02-18-16 09:28]:

...

On Thu, Feb 18, 2016 at 4:23 PM, Patrick Shanahan wrote:

...
https://bugzilla.opensuse.org/show_bug.cgi?id=966535#c3

(In reply to patrick shanahan from comment #2) Dr. Werner Fink 2016-02-15 13:26:21 UTC

...
Perhaps because openSUSE has multi-user.target and requires internet and there is no internet w/o wpa_supplicant. And it's runlevel 3.

Yep ... and both wpa_supplicant.service and wicked services are not part of systemd and therefore this bug does *not* belong to systemd.

It does, although more upstream. Still, it is problem in systemd that affects multiple programs, so suggesting to workaround it in each individual program is not productive.

...
systemd init simply does what the dependency rules do indicate.

There is no way to express dynamic dependencies created by socket activation; and socket activation is marketed as one of systemd advantage. So it should be able to cope with this situation.

Anyway, there is upstream issue opened as noted below. This affects more than one distribution.

It appears that even though this thread was referenced in the openSUSE bug report, it is not being read or followed. Please add your comments to the bug report: https://bugzilla.opensuse.org/show_bug.cfg?id-966535 -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Roman Bysh

20:21

New subject: [opensuse-factory] UPDATE (sorry situation) re: changing runlevel drops internet connection

On 02/18/2016 10:30 AM, Patrick Shanahan wrote:

...

* Andrei Borzenkov [02-18-16 09:28]:

...
On Thu, Feb 18, 2016 at 4:23 PM, Patrick Shanahan wrote:

...
https://bugzilla.opensuse.org/show_bug.cgi?id=966535#c3

(In reply to patrick shanahan from comment #2) Dr. Werner Fink 2016-02-15 13:26:21 UTC

...
Perhaps because openSUSE has multi-user.target and requires internet and there is no internet w/o wpa_supplicant. And it's runlevel 3.

Yep ... and both wpa_supplicant.service and wicked services are not part of systemd and therefore this bug does *not* belong to systemd.

It does, although more upstream. Still, it is problem in systemd that affects multiple programs, so suggesting to workaround it in each individual program is not productive.

...
systemd init simply does what the dependency rules do indicate.

There is no way to express dynamic dependencies created by socket activation; and socket activation is marketed as one of systemd advantage. So it should be able to cope with this situation.

Anyway, there is upstream issue opened as noted below. This affects more than one distribution.

It appears that even though this thread was referenced in the openSUSE bug report, it is not being read or followed. Please add your comments to the bug report: https://bugzilla.opensuse.org/show_bug.cfg?id-966535

The links show Object not found. -- Cheers! Roman IRC: 551368250 ============== -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Yamaban

20:28

New subject: [opensuse-factory] Re: UPDATE (sorry situation) re: changing runlevel drops internet connection

On Thu, 18 Feb 2016 21:21, Roman Bysh wrote:

...

On 02/18/2016 10:30 AM, Patrick Shanahan wrote:

...
* Andrei Borzenkov [02-18-16 09:28]:

...
On Thu, Feb 18, 2016 at 4:23 PM, Patrick Shanahan wrote:

...
https://bugzilla.opensuse.org/show_bug.cgi?id=966535#c3

(In reply to patrick shanahan from comment #2) Dr. Werner Fink 2016-02-15 13:26:21 UTC

...
Perhaps because openSUSE has multi-user.target and requires internet and there is no internet w/o wpa_supplicant. And it's runlevel 3.

Yep ... and both wpa_supplicant.service and wicked services are not part of systemd and therefore this bug does *not* belong to systemd.

It does, although more upstream. Still, it is problem in systemd that affects multiple programs, so suggesting to workaround it in each individual program is not productive.

...
systemd init simply does what the dependency rules do indicate.

There is no way to express dynamic dependencies created by socket activation; and socket activation is marketed as one of systemd advantage. So it should be able to cope with this situation.

Anyway, there is upstream issue opened as noted below. This affects more than one distribution.

It appears that even though this thread was referenced in the openSUSE bug report, it is not being read or followed. Please add your comments to the bug report: https://bugzilla.opensuse.org/show_bug.cfg?id-966535

The links show Object not found.

Wrong char in last link, replace "-" with "=": https://bugzilla.opensuse.org/show_bug.cgi?id=966535 - Yamaban. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Roman Bysh

20:31

New subject: [opensuse-factory] Re: UPDATE (sorry situation) re: changing runlevel drops internet connection

On 02/18/2016 03:28 PM, Yamaban wrote:

...

On Thu, 18 Feb 2016 21:21, Roman Bysh wrote:

...
On 02/18/2016 10:30 AM, Patrick Shanahan wrote:

...
* Andrei Borzenkov [02-18-16 09:28]:

...
On Thu, Feb 18, 2016 at 4:23 PM, Patrick Shanahan wrote:

...
https://bugzilla.opensuse.org/show_bug.cgi?id=966535#c3

(In reply to patrick shanahan from comment #2) Dr. Werner Fink 2016-02-15 13:26:21 UTC

...
Perhaps because openSUSE has multi-user.target and requires internet and there is no internet w/o wpa_supplicant. And it's runlevel 3.

Yep ... and both wpa_supplicant.service and wicked services are not part of systemd and therefore this bug does *not* belong to systemd.

It does, although more upstream. Still, it is problem in systemd that affects multiple programs, so suggesting to workaround it in each individual program is not productive.

...
systemd init simply does what the dependency rules do indicate.

There is no way to express dynamic dependencies created by socket activation; and socket activation is marketed as one of systemd advantage. So it should be able to cope with this situation.

Anyway, there is upstream issue opened as noted below. This affects more than one distribution.

It appears that even though this thread was referenced in the openSUSE bug report, it is not being read or followed. Please add your comments to the bug report: https://bugzilla.opensuse.org/show_bug.cfg?id-966535

The links show Object not found.

Wrong char in last link, replace "-" with "=": https://bugzilla.opensuse.org/show_bug.cgi?id=966535

- Yamaban. Good catch.

Roman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Patrick Shanahan

21:06

New subject: [opensuse-factory] UPDATE (sorry situation) re: changing runlevel drops internet connection

* Patrick Shanahan [02-18-16 10:31]: correct link:

...

It appears that even though this thread was referenced in the openSUSE bug report, it is not being read or followed. Please add your comments to the bug report: https://bugzilla.opensuse.org/show_bug.cfg?id-966535

https://bugzilla.opensuse.org/show_bug.cfg?id=966535 -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

3001

Age (days ago)

3009

Last active (days ago)

List overview

Download

17 comments

4 participants

participants (4)

Andrei Borzenkov
Patrick Shanahan
Roman Bysh
Yamaban