-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 2013-10-22 05:37, Cristian Rodríguez wrote:

El 21/10/13 18:27, Carlos E. R. escribió:

I believe we should implement offline updates http://www.freedesktop.org/wiki/Software/systemd/SystemUpdates/ at some point to bypass this kind of problems completely.

Mmmm... :-? If you tell people that they have to reboot Linux twice to get updates, you will see a lot of grunting and comparisons to Windows. - -- Cheers / Saludos, Carlos E. R. (from 12.3 x86_64 "Dartmouth" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.19 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iEYEARECAAYFAlJmRKMACgkQtTMYHG2NR9XWFwCcD9TwzfPCuBXp0gZfFKz73IU8 bf4An3rsYqGSDG25CuGJu8pLURPoy/H+ =7YA3 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Tuesday 22 October 2013, Cristian Rodríguez wrote:

El 22/10/13 06:25, Carlos E. R. escribió:

...

If you tell people that they have to reboot Linux twice

They do not have to reboot twice..the process is automated..

HAHA selling automatic reboots as a feature ...

to get

...

updates, you will see a lot of grunting and comparisons to Windows.

So what ?

Yeah, I'm sure after the freedesktop guys have finished their job then the usual linux distro will behave even more funny than win98 :) cu, Rudi -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

* Cristian Rodríguez [2013-10-22 17:20]:

El 22/10/13 11:37, Ruediger Meier escribió:

...

On Tuesday 22 October 2013, Cristian Rodríguez wrote:

...

El 22/10/13 06:25, Carlos E. R. escribió:

...

If you tell people that they have to reboot Linux twice

They do not have to reboot twice..the process is automated..

HAHA selling automatic reboots as a feature ...

I am not selling anything, I am however making an statement of fact, that "online" updates do not work under certain scenarios and that I do not want to see a series of endless workarounds added just to make them work. a clean solution is to require offline updates when *certain* components are to be upgraded just like in other OSs.

The design is utterly stupid since it requires two reboots plus downtime for installing the updates while offline instead of just one by installing updates in a snapshot while online. If they're copycatting other OS they could at least copy those which have solved this in a smarter way. -- Guido Berhoerster -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

* Claudio Freire [2013-10-22 18:06]:

On Tue, Oct 22, 2013 at 12:45 PM, Guido Berhoerster wrote:

...

* Cristian Rodríguez [2013-10-22 17:20]:

...

El 22/10/13 11:37, Ruediger Meier escribió:

...

On Tuesday 22 October 2013, Cristian Rodríguez wrote:

...

El 22/10/13 06:25, Carlos E. R. escribió:

...

If you tell people that they have to reboot Linux twice

They do not have to reboot twice..the process is automated..

HAHA selling automatic reboots as a feature ...

I am not selling anything, I am however making an statement of fact, that "online" updates do not work under certain scenarios and that I do not want to see a series of endless workarounds added just to make them work. a clean solution is to require offline updates when *certain* components are to be upgraded just like in other OSs.

The design is utterly stupid since it requires two reboots plus downtime for installing the updates while offline instead of just one by installing updates in a snapshot while online. If they're copycatting other OS they could at least copy those which have solved this in a smarter way.

To be fair, snapshot-capable FS are far from the norm in linux currently.

Well, yeah Linux is lagging behind. But it should be doable with LVM thin provisioning if you don't want to rely on brtfs or ZoL and it's not that systemd can run on older distros/kernels anyway. -- Guido Berhoerster -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Tuesday 22 October 2013, Cristian Rodríguez wrote:

El 22/10/13 11:37, Ruediger Meier escribió:

...

On Tuesday 22 October 2013, Cristian Rodríguez wrote:

...

El 22/10/13 06:25, Carlos E. R. escribió:

...

If you tell people that they have to reboot Linux twice

They do not have to reboot twice..the process is automated..

HAHA selling automatic reboots as a feature ...

I am not selling anything, I am however making an statement of fact, that "online" updates do not work under certain scenarios and that I do not want to see a series of endless workarounds added just to make them work.

Do you have a particular example for such a certain scenario? The issue of this thread is for sure not such a scenario. How on earth one could think that DNS resolving issues are caused by 2 missing reboots? Something else is utterly broken here and need to get fixed.

a clean solution is to require offline updates when *certain* components are to be upgraded just like in other OSs.

The clean solution would be to kick out the whole freedesktop crap. At least would be nice if freedesktop would only break desktop related stuff instead of whole systems. cu, Rudi -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

В Tue, 22 Oct 2013 17:39:06 +0200 Stefan Seyfried пишет:

The only thing where a reboot is warranted is a kernel update.

Restarting your DE is not much different from reboot for most users and reboot has advantage of guaranteeing that everything is really restarted. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 В Tue, 22 Oct 2013 19:57:34 +0200 "Carlos E. R." пишет:

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

On 2013-10-22 19:44, Andrey Borzenkov wrote:

...

В Tue, 22 Oct 2013 17:39:06 +0200 Stefan Seyfried пишет:

...

The only thing where a reboot is warranted is a kernel update.

Restarting your DE is not much different from reboot for most users and reboot has advantage of guaranteeing that everything is really restarted.

True.

I have right now 65 windows opened in my work desktop, not counting those inside two virtual machines.

I do not restart the desktop in weeks, if I can avoid it. I hibernate the machine.

Once I restart the desktop, restarting the rest with a reboot is not that much.

Ah, not only the kernel requires a reboot: a glibc update does, to. Or a systemd update (PID 1)

Actually systemd is capable of re-executing itself which is done on update. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.19 (GNU/Linux) iEYEARECAAYFAlJmvdsACgkQR6LMutpd94zkswCghlNztuh2nUvzDCIUyPAosKXw kMwAn1gYdUT/I1n4fgvQULTQIwd62Mxg =13Im -----END PGP SIGNATURE----- N▀╖╡ФЛr╦⌡yИ ┼Z)z{.╠Гзrз+кК╖╡ФЛr╦⌡z┼^·к╛z┼ЮN┤(·ж°╤ь^ё ч╜И ┼Z)z{.╠Гзrз+кЙ0²ЙХ╔ИЛ╨г╗╝

El 22/10/13 14:57, Carlos E. R. escribió:

Ah, not only the kernel requires a reboot: a glibc update does, to. Or a systemd update (PID 1)

systemctl daemon-reexec avoids that.. however it is indeed true that a few other components require reboot for the update to be complete. -- "If debugging is the process of removing bugs, then programming must be the process of putting them in." - Edsger Dijkstra -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

El 22/10/13 15:16, Ruediger Meier escribió:

For example?

Usually dbus or any other patch that is marked with the "Reboot Required: yes" . -- "If debugging is the process of removing bugs, then programming must be the process of putting them in." - Edsger Dijkstra -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 2013-10-22 20:05, Cristian Rodríguez wrote:

El 22/10/13 14:57, Carlos E. R. escribió:

...

Ah, not only the kernel requires a reboot: a glibc update does, to. Or a systemd update (PID 1)

systemctl daemon-reexec avoids that.. however it is indeed true that a few other components require reboot for the update to be complete.

It is not automatic - just now I updated a VM using 12.3:

The following running processes use deleted files:

PID | PPID | UID | Login | Command | Service | Files ------+-------+------+------------+----------------------------+--------------+------------------------------------- 258 | 1 | 0 | root | systemd-journald (deleted) | | /usr/lib64/libudev.so.1.1.6 | | | | | | /usr/lib64/libsystemd-daemon.so.0.-> | | | | | | /usr/lib/systemd/systemd-journald ->

-- Cheers / Saludos, Carlos E. R. (from 12.3 x86_64 "Dartmouth" at Telcontar)

On 2013-10-22 20:26, Cristian Rodríguez wrote:

Did you read what you just posted.. ? you are confusing systemd (PID1) with the systemd journal.. totally different things..

Ok... Nevertheless, how do I restart that? -- Cheers / Saludos, Carlos E. R. (from 12.3 x86_64 "Dartmouth" at Telcontar)

On Tue, 22 Oct 2013 20:19, Carlos E. R. wrote:

On 2013-10-22 20:05, Cristian Rodríguez wrote:

...

El 22/10/13 14:57, Carlos E. R. escribió:

...

Ah, not only the kernel requires a reboot: a glibc update does, to. Or a systemd update (PID 1)

systemctl daemon-reexec avoids that.. however it is indeed true that a few other components require reboot for the update to be complete.

It is not automatic - just now I updated a VM using 12.3:

...

The following running processes use deleted files:

PID | PPID | UID | Login | Command | Service | Files ------+-------+------+------------+----------------------------+--------------+------------------------------------- 258 | 1 | 0 | root | systemd-journald (deleted) | | /usr/lib64/libudev.so.1.1.6 | | | | | | /usr/lib64/libsystemd-daemon.so.0.-> | | | | | | /usr/lib/systemd/systemd-journald ->

Huh? This works just fine: root:/# systemctl restart systemd-journald But, yes, DBus and NetworkManager do NOT play nice with each other. Maybe a "restart networkmanager" should be in %post for dbus package. - Yamaban.

On Tue, 22 Oct 2013 20:38, Cristian Rodríguez wrote:

El 22/10/13 15:35, Yamaban escribió:

...

Maybe a "restart networkmanager" should be in %post for dbus package.

No, that will make matters worst.. also package X %post install scripts should only deal with processes or files that belong to package X.. not to Y or Z.

Then use your position to get upstream networkmanager to play nice with dbus, and thus SOLVE the root of this misbehavior (bug). openSUSE favours networkmanager as network handler even for wired network. IMHO networkmanager is mostly crap, maybe better crap than others, but still. Do not get me started on the GUI situation, "shit" doesn't even start. (How many still open bugs?) - Yamaban.

On 2013-10-22 20:35, Yamaban wrote:

On Tue, 22 Oct 2013 20:19, Carlos E. R. wrote:

root:/# systemctl restart systemd-journald

Ok... The update would also need to restart "gdbus". Too much. -- Cheers / Saludos, Carlos E. R. (from 12.3 x86_64 "Dartmouth" at Telcontar)

Am 22.10.2013 19:57, schrieb Carlos E. R.:

Ah, not only the kernel requires a reboot: a glibc update does, to. Or a systemd update (PID 1)

No. Running processes use the old glibc, but that usually does not matter. If it does (means: if there was a really critical remote exploit in glibc), I'd reboot anyway. But when have you last seen such a bug? I'm updating my factory machine all the time, whenever new packages are available. But I reboot only when there is a significantly newer kernel version or a particular kernel fix I'd like to have. -- Stefan Seyfried "If your lighter runs out of fluid or flint and stops making fire, and you can't be bothered to figure out about lighter fluid or flint, that is not Zippo's fault." -- bkw -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

В Tue, 22 Oct 2013 21:06:18 +0200 Stefan Seyfried пишет:

Am 22.10.2013 19:44, schrieb Andrey Borzenkov:

...

В Tue, 22 Oct 2013 17:39:06 +0200 Stefan Seyfried пишет:

...

The only thing where a reboot is warranted is a kernel update.

Restarting your DE is not much different from reboot for most users

That's why I also never do that, the DE gets restarted after reboot.

So what?

Any update that touches your running DE requires restarting DE. And there are much more updates that touch DE than kernel updates. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 2013-10-23 06:27, Stefan Seyfried wrote:

Am 23.10.2013 04:24, schrieb Andrey Borzenkov:

...

Any update that touches your running DE requires restarting DE.

No. Sorry, that's just not true. I sometimes restart xfwm or xfdesktop manually, but that's totally nondestructive. No need to log out. Even the running yakuake instance survives a kdelibs update totally fine, it just does not use the newest versions.

Not using the newest version is the same as not updating. The update will happen only when you restart it. -- Cheers / Saludos, Carlos E. R. (from 12.3 x86_64 "Dartmouth" at Telcontar)

* Michal Vyskocil [10-24-13 08:28]: [...]

vi /etc/zypp/zypp.conf

commit.downloadMode = DownloadInAdvance

I already thought this is a default, as I don't have anything in my zypp, but I am sure it does work like that.

er, commit.downloadMode = DownloadInHeaps ?? -- (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

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 2013-10-22 14:23, Greg Freemyer wrote:

"Carlos E. R." <> wrote:

...

Bug somewhere?

What was the opensuse version?

Look above in the subject, please ;-)

Old zypper would download one package at a time and install it and could have that problem.

Newer zypper first downloads all packages first, then starts upgrading. Even if a package breaks the network it should not keep zypper from completing.

Well, I was using YaST. I said so. - -- Cheers / Saludos, Carlos E. R. (from 12.3 x86_64 "Dartmouth" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.19 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iEYEARECAAYFAlJmd64ACgkQtTMYHG2NR9WiDgCfX6XJ3TVNu1leFvtqu4MDY5SI XBkAnit1mcoBnaAN/G8FL72V4SOk3o7/ =ZgRn -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 2013-10-22 17:26, Greg Freemyer wrote:

On Tue, Oct 22, 2013 at 9:03 AM, Carlos E. R. <> wrote:

Well, YaST had been bitrotting for years prior to 13.1. I don't know if it got much actual improvement in 13.1, or just a conversion to ruby.

About just the conversion.

Sounds like you should create a feature request in fate to have it do a full set of package downloads first, then start doing installs of the packages.

I believe it can do so, but has to be configured somewhere. The update I did today in the VM downloaded many many packages before it started to install any of them. I believe it does so in bunches. And it was not disrupted, but network there is ifup and cable.

Then create a bugzilla for 13.1 that shows the failure and points to the FATE entry as a possible fix.

I believe it is more a single package failure. The rpm should contain a script to restart the services it disrupted. I could write a bugzilla, yes, but I do not know if it is logged or in which file, and what part of the file. I'll try to find out. - -- Cheers / Saludos, Carlos E. R. (from 12.3 x86_64 "Dartmouth" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.19 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iEYEARECAAYFAlJmu0AACgkQtTMYHG2NR9U0FwCeKrAALICryKfUW85NW356aBO3 kv8AnRThUtVofZfnmjAruCOhfdA0Fslf =jeHY -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 10/22/2013 07:52 PM, Carlos E. R. wrote:

...

Then create a bugzilla for 13.1 that shows the failure and points to the FATE entry as a possible fix.

I believe it is more a single package failure. The rpm should contain a script to restart the services it disrupted. I could write a bugzilla, yes, but I do not know if it is logged or in which file, and what part of the file. I'll try to find out.

Any update on the bugzilla report? If there were any Yast errors, you would find them in /var/log/YaST2/y2log Some errors or addtional output from RPMs can be found at /var/log/zypp/history Anyway, I also suspect some other package than Yast, zypp/history log could tell you more as you have the last modification time. Thx && Bye Lukas -- Lukas Ocilka, Cloud & Systems Management Department SUSE LINUX s.r.o., Praha -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Tuesday 2013-10-22 15:03, Carlos E. R. wrote:

...

Old zypper would download one package at a time and install it and could have that problem.

[aka DownloadAsNeeded]

...

Newer zypper first downloads all packages first, then starts upgrading. Even if a package breaks the network it should not keep zypper from completing.

[aka DownloadInAdvance]

Well, I was using YaST. I said so.

But even yast still uses libzypp, and DownloadAsNeeded and DownloadInAdvance are zypp.conf options. Of course, it is still possible that yast works around the setting found in zypp.conf, but that would actually be doubly worth a bug report :) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org