Felix Miata writes:

I think you are right. NTP has apparently been deprecated. I've been doing:

I actually use it to provide a stratum-1 network time source based on a GPS PPS, so timesyncd (which like all SNTP clients you shouldn't be using either if you care about time) isn't cutting it. Regards, Achim. -- +<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+ SD adaptations for KORG EX-800 and Poly-800MkII V0.9: http://Synth.Stromeko.net/Downloads.html#KorgSDada -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Brüns, Stefan wrote:

timesyncd is completely sufficient for most desktop users who only need their clock to be accurate in the millisecond range.

Are you really trusting it to better than say 25ms? (I'm running it on the laptop, too. But not on any machine that needs single-digit ms accuracy...) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Mittwoch, 14. November 2018 21:00:20 CET Achim Gratz wrote:

Brüns, Stefan writes:

...

NTPd has been replaced by chrony.

Please, I have asked about an apparent problem with systemd integration for NTPd, not unsolicited (and often hilariously wrong) advice about other time services.

Chrony very likely is a suitable replacement for NTPd, so using chrony instead will avoid the problems you have with NTPd. Do you have any requirements beyond GPS PPS you forgot to mention? As ntpd no longer is the default ntp server implementation for (open)SUSE, it is likely more issues appear over the course of time. No reason to be rude ... Regards, Stefan -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Donnerstag, 15. November 2018 09:36:00 CET Stefan Seyfried wrote:

Am 14.11.18 um 22:06 schrieb Brüns, Stefan:

...

Chrony very likely is a suitable replacement for NTPd,

If I'd only find out how to connect my serial DCF77 clock to chrony, this might become true.

The very last paragraph in the comparison I linked previously, ntp-refclock is mentioned: https://github.com/mlichvar/ntp-refclock Kind regards, Stefan -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Samstag, 17. November 2018 13:00:10 CET Stefan Seyfried wrote:

Am 15.11.18 um 16:53 schrieb Brüns, Stefan:

...

On Donnerstag, 15. November 2018 09:36:00 CET Stefan Seyfried wrote:

...

Am 14.11.18 um 22:06 schrieb Brüns, Stefan:

...

Chrony very likely is a suitable replacement for NTPd,

If I'd only find out how to connect my serial DCF77 clock to chrony, this might become true.

The very last paragraph in the comparison I linked previously, ntp-refclock is mentioned:

https://github.com/mlichvar/ntp-refclock

OK, so I should replace ntpd with ntpd plus chrony?

No. ntp-refclock *reuses* hardware drivers distributed as part of ntp and makes them work standalone. It is a very tiny part of ntp. Of course someone could just do a fork of ntp with only the drivers and delete the ntp daemon itself. That would be a win for everybody, nobody would cry "NIH", or blame someone for segmentation. NOT! The NTP *core* is showing its age, so it makes sense to replace it, but reuse as much as possible. Regards, Stefan -- Stefan Brüns / Bergstraße 21 / 52062 Aachen home: +49 241 53809034 mobile: +49 151 50412019

On Samstag, 17. November 2018 19:34:53 CET Stefan Seyfried wrote:

Am 17.11.18 um 17:23 schrieb Stefan Brüns:

...

On Samstag, 17. November 2018 13:00:10 CET Stefan Seyfried wrote:

...

Am 15.11.18 um 16:53 schrieb Brüns, Stefan:

...

On Donnerstag, 15. November 2018 09:36:00 CET Stefan Seyfried wrote:

...

Am 14.11.18 um 22:06 schrieb Brüns, Stefan:

...

Chrony very likely is a suitable replacement for NTPd,

If I'd only find out how to connect my serial DCF77 clock to chrony, this might become true.

The very last paragraph in the comparison I linked previously, ntp-refclock is mentioned:

https://github.com/mlichvar/ntp-refclock

OK, so I should replace ntpd with ntpd plus chrony?

No. ntp-refclock *reuses* hardware drivers distributed as part of ntp and makes them work standalone. It is a very tiny part of ntp.

OK, and then I need a "ntp-broadcast" deamon in addition to chrony to use broadcast / multicast ntp etc.,

Chrony can broadcast. I does not listen to broadcasts, see https:// chrony.tuxfamily.org/ faq.html#_can_code_chronyd_code_be_driven_from_broadcast_multicast_ntp_servers for the reasoning. How many clients do you have in your network? 10.000? 100.000? Or is this just a strawman? Using a GPS PPS clock and broadcast mode is contradicting ... Kind regards, Stefan -- Stefan Brüns / Bergstraße 21 / 52062 Aachen home: +49 241 53809034 mobile: +49 151 50412019

Stefan Seyfried wrote:

Am 17.11.18 um 19:47 schrieb Stefan Brüns:

...

On Samstag, 17. November 2018 19:34:53 CET Stefan Seyfried wrote:

...

...

No. ntp-refclock *reuses* hardware drivers distributed as part of ntp and makes them work standalone. It is a very tiny part of ntp.

OK, and then I need a "ntp-broadcast" deamon in addition to chrony to use broadcast / multicast ntp etc.,

Chrony can broadcast. I does not listen to broadcasts, see https:// chrony.tuxfamily.org/

faq.html#_can_code_chronyd_code_be_driven_from_broadcast_multicast_ntp_servers

...

for the reasoning.

How many clients do you have in your network? 10.000? 100.000?

about 20, but ntp broadcast is just an easy no-configuration-needed setup that I'm using now since about 15 years.

FWIW, exactly the same config here, including the DCF77 serial clock. -- Per Jessen, Zürich (2.1°C) member, openSUSE Heroes. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Achim Gratz composed on 2018-11-13 21:52 (UTC+0100):

Felix Miata writes:

...

I think you are right. NTP has apparently been deprecated. I've been doing:

I actually use it to provide a stratum-1 network time source based on a GPS PPS, so timesyncd (which like all SNTP clients you shouldn't be using either if you care about time) isn't cutting it.

I forgot to mention why "I've been doing" (the switch to systemd-timesyncd) was because of the ntp shutdown delays, which I mentioned on the opensuse list, without getting a solution, last May: https://lists.opensuse.org/opensuse/2018-05/msg00912.html Failure to get a solution repeated in July: https://lists.opensuse.org/opensuse-support/2018-07/msg00003.html I only discovered systemd-timesyncd around two weeks ago. -- Evolution as taught in public schools is religion, not science. Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

14.11.2018 23:05, Achim Gratz пишет:

Per Jessen writes:

...

According to the manpage(s), when there is no specific ExecStop action, the process will just be sent a SIGTERM.

That's my assumption, too. But obviously systemd doesn't think the process has terminated

You asked systemd and it told you so privately? May be process really has not terminated? Do you have any evidence that at 21:44:55 ntpd processes were stopped and so it was systemd fault to not notice it?

and keeps waiting for that to happen:

If you can reproduce it reliably, enable systemd debug output before shutdown (/bin/kill -RTMIN+22 1) and check whether (and when) ntpd processes terminated.

Nov 13 21:54:55 Gertrud systemd[1]: Stopping Login Service... Nov 13 21:54:55 Gertrud systemd[1]: Stopped target System Time Synchronized. Nov 13 21:54:55 Gertrud systemd[1]: Stopping NTP Server Daemon... Nov 13 21:54:55 Gertrud systemd[1]: Stopping Permit User Sessions... Nov 13 21:54:55 Gertrud systemd[1]: Received SIGRTMIN+20 from PID 17247 (plymouthd). Nov 13 21:54:55 Gertrud kernel: pps pps0: removed Nov 13 21:54:55 Gertrud systemd[1]: Stopped Permit User Sessions. Nov 13 21:54:55 Gertrud systemd[1]: Stopped target Remote File Systems. Nov 13 21:54:55 Gertrud systemd[1]: Stopping Login and scanning of iSCSI devices... Nov 13 21:54:55 Gertrud systemd[1]: Stopped target Remote File Systems (Pre). Nov 13 21:54:55 Gertrud systemd[1]: Started Show Plymouth Power Off Screen. Nov 13 21:54:55 Gertrud iscsiadm[17249]: iscsiadm: No matching sessions found Nov 13 21:54:55 Gertrud iscsiadm[17250]: iscsiadm: No matching sessions found Nov 13 21:54:55 Gertrud systemd[1]: Stopped Login and scanning of iSCSI devices. Nov 13 21:54:55 Gertrud systemd[1]: Stopped Login Service. Nov 13 21:54:55 Gertrud systemd[1]: Stopped target User and Group Name Lookups. Nov 13 21:54:55 Gertrud systemd[1]: Stopped Restore /run/initramfs on shutdown. Nov 13 21:55:57 Gertrud systemd[1]: Stopped NTP Server Daemon.

Regards, Achim.

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

Achim Gratz writes:

So it seems that indeed the asynchronous resolver is what hangs, while ntpd itself has already exited. Why it does that is not clear, but I have a hunch… one of my rasPi shredded its SD card recently, so I took it out of service until I get a replacement. It's still resolved in DNS, but of course won't answer. I have removed that server from the config, let's see if that helps (a quick restart went smoothly, but that may not mean much).

The peer config for the non-answering rasPi was indeed the reason for hanging up the async resolver process on exit. Regards, Achim. -- +<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+ Wavetables for the Waldorf Blofeld: http://Synth.Stromeko.net/Downloads.html#BlofeldUserWavetables -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

15.11.2018 2:34, Martin Wilck пишет:

On Tue, 2018-11-13 at 21:24 +0100, Achim Gratz wrote:

...

The handling of ntpd apparently switched from an old-style rc script to a systemd unit. Since then the stop script will take somewhere between ten seconds to a minute to declare success and continue with the shutdown.

Strange - AFAICS the systemd unit files for ntp have existed for more than 4 years in factory.

This is not the first report that ntpd service takes long time on shutdown.

...

I don't see why it should take that long to shut down since ntpd really just closes the ntp logs and exits. The systemd unit file treats it as a forking daemon, which probably makes the check of whether the daemon has exited more involved.

Not really involved. For forking daemons, systemd simply kills and checks the PID mentioned in the PIDfile directive.

systemd waits for all processes to stop. ntpd starts at least one more process for asynchronous resolver. It's possible that this process may get blocked and ignores SIGTERM. It is even possible that it gets blocked somewhere in a library call and so the reason is outside of ntpd. I am not sure how to debug it, given that it seems to happen on shutdown only (and logs presented here are obviously from system shutdown as well). I thought that it may be related to network being stopped too early, but simple minded "systemctl stop network; systemctl stop ntpd" still has no delay. But if someone can reproduce it with high probability (I cannot) running with systemd debug level output is the first step.

...

But ntpd can easily be told not to fork, so I think it should use the "-n" option in ExecStart and all

systemd does exactly the same in both cases - it either runs ExecStop or sends signal, so I do not see how it is going to change anything.

...

the stuff that the current start script is doing should go into ExecStartPre?

Please consider opening a bug.

Martin

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

On 18/11/2018 14.51, Marco Calistri wrote:

...

I was relying on ntpd service since ever until I read about "chronyd" as the substitute of ntpd adopted by new opensuse releases, then I switched to chronyd.

Despite this it seems ntp package being still necessary (?) or could be removed by the system?

To verify if my system is synchronized I use the command "timedatectl":

marco@linux-turion64:~> timedatectl Local time: dom 2018-11-18 11:40:48 -02 Universal time: dom 2018-11-18 13:40:48 UTC RTC time: dom 2018-11-18 11:40:48 Time zone: America/Sao_Paulo (-02, -0200) System clock synchronized: yes NTP service: inactive RTC in local TZ: yes

Interesting command! I did not know about it. But it would be rather more interesting if it displayed milliseconds, as it does it is not sufficient. It also, now that I think, does not display the network time at all, in my case from ntp, so I do not know how well my clock is synced - or not. -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)

On 18/11/2018 20.27, Cristian Rodríguez wrote:

El 18-11-2018 a las 15:56, Carlos E. R. escribió:

...

Interesting command! I did not know about it. But it would be rather more interesting if it displayed milliseconds, as it does it is not sufficient.

It also, now that I think, does not display the network time at all,

Of course, it displays the system time.. not the network time.

...

in my case from ntp, so I do not know how well my clock is synced - or not.

"NTP service:" will always be "inactive" if you are running anything but systemd-timesyncd.. it no longer knows about the status of other implementations

But I don't. All my systems are still using ntpd, except my new and tiny laptop which uses chrony :-) I do not know yet what advantages chrony brings. My laptops might use systemd-timesyncd, but I have not changed any of them yet, not seeing the advantage either. -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)

On 19/11/2018 14.09, Cristian Rodríguez wrote:

El 18-11-2018 a las 16:52, Carlos E. R. escribió:

...

I do not know yet what advantages chrony brings.

Let's say the traditional ntp daemon has not aged very well.. and is steadly bitrotting. until  ntimed, which is the reference implementation replacement is mature enough, chrony is probably the best alternative if you need an  NTP *server* or a very high precision client. otherwise, just use systemd-timesyncd.

Well, it seems openSUSE 15.0 defaults to chrony, that's what my small laptop got on fresh install. If it uses less memory, I'm certainly very interested. How does it cope if there is no internet? I have said often that the installation should ask what type of computer it is, and if it is a laptop, it should change things, like installing systemd-timesyncd instead. -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)

...

I have said often that the installation should ask what type of computer it is, and if it is a laptop, it should change things, like installing systemd-timesyncd instead.

Normally a laptop does have a battery to keep the clock running [...]

However when such a system does not have such a clock [...]

This is also a nice example why every installer component needs an expert mode: There are always exceptions. Here, a laptop's cmos battery may be empty, e.g. due to age. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Op maandag 19 november 2018 20:00:23 CET schreef Carlos E. R.:

Yes. But previous tools took notice on how much the CMOS clock drifted while the computer was off, before setting up the system time. Similarly, the frequency of the system clock was adjusted.

Does systemd-timesyncd take this into account?

The system time is stored regularly. When systemd-timesyncd is executed early in the boot process, the clock is restored at the time as can be seen in the journal. Obviously there is no indication of how well the system time changes. It changes in the same rate by which it changes between synchronization with the Internet clock before the boot. Later when an Internet connection is established the clock changes again to the world clock. After that it changes by some clock device in the processor until it is synchronized again with the world clock. -- fr.gr. member openSUSE Freek de Kruijf -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Wed, 21 Nov 2018 13:33:56 +0100, "Carlos E. R." wrote:

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

On 19/11/2018 15.19, Carlos E. R. wrote:

...

On 19/11/2018 14.09, Cristian Rodríguez wrote:>> El 18-11-2018 a las 16:52, Carlos E. R. escribió:

...

[...]

...

Let's say the traditional ntp daemon has not aged very well.. and is steadly bitrotting. until ntimed, which is the reference implementation replacement is mature enough, chrony is probably the best alternative if you need an NTP *server* or a very high precision client. otherwise, just use systemd-timesyncd.

Well, it seems openSUSE 15.0 defaults to chrony, that's what my small laptop got on fresh install.

If it uses less memory, I'm certainly very interested.

chrony:

top - 13:21:03 up 13 days, 12:59, 3 users, load average: 0.23, 0.46, 0.54 Tasks: 342 total, 1 running, 341 sleeping, 0 stopped, 0 zombie %Cpu(s): 17.9 us, 4.7 sy, 0.1 ni, 76.7 id, 0.1 wa, 0.0 hi, 0.5 si, 0.0 st KiB Mem : 3942200 total, 1096244 free, 2397364 used, 448592 buff/cache KiB Swap: 12582908 total, 9669852 free, 2913056 used. 1185596 avail Mem

PID USER PR NI VIRT RES SHR SWAP S %CPU %MEM TIME+ COMMAND 1166 chrony 20 0 90388 76 0 132 S 0.000 0.002 0:00.72 chronyd

systemd-timesyncd

top - 13:25:34 up 13 days, 13:03, 3 users, load average: 0.39, 0.60, 0.59 Tasks: 323 total, 1 running, 322 sleeping, 0 stopped, 0 zombie %Cpu(s): 17.9 us, 4.7 sy, 0.1 ni, 76.7 id, 0.1 wa, 0.0 hi, 0.5 si, 0.0 st KiB Mem : 3942200 total, 1002344 free, 2444456 used, 495400 buff/cache KiB Swap: 12582908 total, 9683932 free, 2898976 used. 1113416 avail Mem

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 14070 systemd+ 20 0 133332 3024 2560 S 0.000 0.077 0:00.03 systemd-timesyn

Well, no, it seems systemd-timesyncd uses much more memory.

I'm going back to chrony.

If size is the only argument, ntpd is the smallest USER PID PPID PRI NI SIZE STIME TTY TIME COMMAND ntp 1533 1 20 0 9962 14:48:24 - 0:00:05 /usr/sbin/ntpd -p /var/run/ntp/ntpd.pid -g -u ntp:ntp -c /etc/ntp.conf ntp 1538 1533 20 0 9962 14:48:24 - 0:00:00 ntpd: asynchronous dns resolver USER PID PPID PRI NI SIZE STIME TTY TIME COMMAND chrony 346 1 20 0 22305 13:43:27 - 0:00:00 /usr/sbin/chronyd PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 1534 ntp 20 0 39848 4080 3300 S 0.000 0.025 0:00.00 ntpd 1535 ntp 20 0 39848 776 0 S 0.000 0.005 0:00.00 ntpd PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 346 chrony 20 0 89220 2484 2300 S 0.000 0.015 0:00.01 chronyd -- H.Merijn Brand http://tux.nl Perl Monger http://amsterdam.pm.org/ using perl5.00307 .. 5.29 porting perl5 on HP-UX, AIX, and openSUSE http://mirrors.develooper.com/hpux/ http://www.test-smoke.org/ http://qa.perl.org http://www.goldmark.org/jeff/stupid-disclaimers/

On Wed, 21 Nov 2018 14:07:13 +0100, "Carlos E. R." wrote:

On 21/11/2018 13.51, H.Merijn Brand wrote:

...

On Wed, 21 Nov 2018 13:33:56 +0100, "Carlos E. R." <> wrote:

[...]

...

...

chrony:

top - 13:21:03 up 13 days, 12:59, 3 users, load average: 0.23, 0.46, 0.54 Tasks: 342 total, 1 running, 341 sleeping, 0 stopped, 0 zombie %Cpu(s): 17.9 us, 4.7 sy, 0.1 ni, 76.7 id, 0.1 wa, 0.0 hi, 0.5 si, 0.0 st KiB Mem : 3942200 total, 1096244 free, 2397364 used, 448592 buff/cache KiB Swap: 12582908 total, 9669852 free, 2913056 used. 1185596 avail Mem

PID USER PR NI VIRT RES SHR SWAP S %CPU %MEM TIME+ COMMAND 1166 chrony 20 0 90388 76 0 132 S 0.000 0.002 0:00.72 chronyd

systemd-timesyncd

top - 13:25:34 up 13 days, 13:03, 3 users, load average: 0.39, 0.60, 0.59 Tasks: 323 total, 1 running, 322 sleeping, 0 stopped, 0 zombie %Cpu(s): 17.9 us, 4.7 sy, 0.1 ni, 76.7 id, 0.1 wa, 0.0 hi, 0.5 si, 0.0 st KiB Mem : 3942200 total, 1002344 free, 2444456 used, 495400 buff/cache KiB Swap: 12582908 total, 9683932 free, 2898976 used. 1113416 avail Mem

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 14070 systemd+ 20 0 133332 3024 2560 S 0.000 0.077 0:00.03 systemd-timesyn

Well, no, it seems systemd-timesyncd uses much more memory.

I'm going back to chrony.

If size is the only argument, ntpd is the smallest

USER PID PPID PRI NI SIZE STIME TTY TIME COMMAND ntp 1533 1 20 0 9962 14:48:24 - 0:00:05 /usr/sbin/ntpd -p /var/run/ntp/ntpd.pid -g -u ntp:ntp -c /etc/ntp.conf ntp 1538 1533 20 0 9962 14:48:24 - 0:00:00 ntpd: asynchronous dns resolver

USER PID PPID PRI NI SIZE STIME TTY TIME COMMAND chrony 346 1 20 0 22305 13:43:27 - 0:00:00 /usr/sbin/chronyd

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 1534 ntp 20 0 39848 4080 3300 S 0.000 0.025 0:00.00 ntpd 1535 ntp 20 0 39848 776 0 S 0.000 0.005 0:00.00 ntpd

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 346 chrony 20 0 89220 2484 2300 S 0.000 0.015 0:00.01 chronyd

Chrony uses half the resident size than ntpd (top output).

But more than twice virtual memory

The other output is from "ps", I guess; I don't know there what "SIZE" means, what exact concept. :-?

The SIZE and RSS fields don't count some parts of a process including the page tables, kernel stack, struct thread_info, and struct task_struct. This is usually at least 20 KiB of memory that is always resident. SIZE is the virtual size of the process (code+data+stack). -- H.Merijn Brand http://tux.nl Perl Monger http://amsterdam.pm.org/ using perl5.00307 .. 5.29 porting perl5 on HP-UX, AIX, and openSUSE http://mirrors.develooper.com/hpux/ http://www.test-smoke.org/ http://qa.perl.org http://www.goldmark.org/jeff/stupid-disclaimers/

On 2018-11-21 14:32, Carlos E. R. wrote:

...

...

The other output is from "ps", I guess; I don't know there what "SIZE" means, what exact concept. :-? The SIZE and RSS fields don't count some parts of a process including the page tables, kernel stack, struct thread_info, and struct task_struct. This is usually at least 20 KiB of memory that is always resident. SIZE is the virtual size of the process (code+data+stack). Ah, it counts virtual size too. Then it is not useful for this comparison.

I have to check some home made things from time to time. Top and ps is fine but when I need to dig deeper i turn to procfs. /proc/[PID]/stat /proc/[PID]/statm /proc/[PID]/status Status is probably the most readable for humans :) egrep -i -e ^rss -e ^vm /proc/[PID]/status And then check "man procfs" and section "/proc/[pid]/status" Cheers, -- /bengan -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 22/11/2018 02:24, Marco Calistri wrote:

Il 21/11/2018 11:32, Carlos E. R. ha scritto:

...

On 21/11/2018 14.21, H.Merijn Brand wrote:

...

On Wed, 21 Nov 2018 14:07:13 +0100, "Carlos E. R." <> wrote:

...

...

Chrony uses half the resident size than ntpd (top output). But more than twice virtual memory Yes, but as Stefan Seyfried says, that is not relevant, it is not occupy ram chips.

...

...

The other output is from "ps", I guess; I don't know there what "SIZE" means, what exact concept. :-? The SIZE and RSS fields don't count some parts of a process including the page tables, kernel stack, struct thread_info, and struct task_struct. This is usually at least 20 KiB of memory that is always resident. SIZE is the virtual size of the process (code+data+stack). Ah, it counts virtual size too. Then it is not useful for this comparison.

Carlos,

Be warned that "somebody" could complain about the fact that you are taking the space of the factory list for "not pretty development topics" :-)

Cheers,

I will be that somebody, guys this is the third or forth thread in a row where the same group of people have caused a thread of 10+ emails within a day that are not related to the development of openSUSE please stop. Consider this your last warning before I escalate it to the board. -- Simon Lees (Simotek) http://simotek.net Emergency Update Team keybase.io/simotek SUSE Linux Adelaide Australia, UTC+10:30 GPG Fingerprint: 5B87 DB9D 88DC F606 E489 CEC5 0922 C246 02F0 014B -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Mittwoch, 21. November 2018 23:24:05 CET Carlos E. R. wrote:

On 21/11/2018 23.19, Simon Lees wrote:

...

On 22/11/2018 02:24, Marco Calistri wrote:

...

Il 21/11/2018 11:32, Carlos E. R. ha scritto:

Carlos,

Be warned that "somebody" could complain about the fact that you are taking the space of the factory list for "not pretty development topics"

:-)

Cheers,

I will be that somebody, guys this is the third or forth thread in a row where the same group of people have caused a thread of 10+ emails within a day that are not related to the development of openSUSE please stop. Consider this your last warning before I escalate it to the board.

I nice and polite reminder is enough. Please do not menace.

If reminding twice a week apparently does not cut it, and some of the posters even show an attitude of - "I know I am off-topic, but I don't care", I would understand some lack of politeness, which I can *not* see here. Simon politely asked to stop, and gave a warning what will happen *iff* the misbehaviour continues. Whoever shows a lack of respect for others will have to live with the consequences. Regards, Stefan-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

* Carlos E. R. [11-22-18 07:43]:

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

On 22/11/2018 00.06, Brüns, Stefan wrote:

...

On Mittwoch, 21. November 2018 23:24:05 CET Carlos E. R. wrote:

...

On 21/11/2018 23.19, Simon Lees wrote:

...

On 22/11/2018 02:24, Marco Calistri wrote:

...

Il 21/11/2018 11:32, Carlos E. R. ha scritto:

Carlos,

Be warned that "somebody" could complain about the fact that you are taking the space of the factory list for "not pretty development topics"

:-)

Cheers,

I will be that somebody, guys this is the third or forth thread in a row where the same group of people have caused a thread of 10+ emails within a day that are not related to the development of openSUSE please stop. Consider this your last warning before I escalate it to the board.

I nice and polite reminder is enough. Please do not menace.

If reminding twice a week apparently does not cut it, and some of the posters even show an attitude of - "I know I am off-topic, but I don't care", I would understand some lack of politeness, which I can *not* see here.

Simon politely asked to stop, and gave a warning what will happen *iff* the misbehaviour continues. Whoever shows a lack of respect for others will have to live with the consequences.

Marco asked politely. I stopped this thread when I saw his post. Simon's was redundant and aggressive.

I'm not aware of other threads. As I do not detect what you consider offtopic or excessive chat here, I ask a *kind* request on every such thread. Usually Patrick does that.

yet the "Thread" continues and is still off-topic here. -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri Registered Linux User #207535 @ http://linuxcounter.net Photos: http://wahoo.no-ip.org/piwigo paka @ IRCnet freenode -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Thu, 2018-11-22 at 15:15 +0100, Manfred Hollstein wrote:

Hi Patrick,

On Thu, 22 Nov 2018, 14:46:55 +0100, Patrick Shanahan wrote:

...

* Carlos E. R. [11-22-18 07:43]:

...

[...] Marco asked politely. I stopped this thread when I saw his post. Simon's was redundant and aggressive.

I'm not aware of other threads. As I do not detect what you consider offtopic or excessive chat here, I ask a *kind* request on every such thread. Usually Patrick does that.

yet the "Thread" continues and is still off-topic here.

if *you* want to complain about people following up on this thread here, why don't you re-direct them using the usual ways to other media? One way would be to *not* post to the same list...

To be honest, *iff* you/other board members believe, we need a pure, development only list, then *create one*!

We have, its this one, most people seem to understand that fine you just need to update the mapping in your head so that factory == development, since the dawn of time openSUSE development has happened in the factory project / repository. Yes some people don't get the name at first so we remind them, but in our opinion its not worth the effort to rename / move this list when most people get it anyway. What openSUSE mailing lists are used for is decided collectively by the openSUSE project not by what someone thinks they should be.

Sorry, but appr. 50% of this thread is mostly dealing with complaints about messages which don't belong here. If the board believes decisions about "ntpd is too old" vs. "chrony is new enough" (because bug reports still come in), then OK, please blame me and potentially black-list me, but I'm really over it now (TM).

Well this thread started with what probably should have been a bug report, but unless someone is consistently posting here rather then making bug reports we will generally let it slide. The rest of the thread might have made sense if we were having a discussion on why we should change the system default rather then what makes sense atm in someones personal case. For reference we are currently following what SLE is / will be using as it means we get enterprise support for free along with additional testing / QA, if someone put together a really compelling argument about why we should not be following SLE for the system default then lets have that discussion. If you just want to discuss the merits of each for your use case then we have the opensuse@ and opensuse-offtopic@ mailing lists, if you need help configuring one of the services we also have opensuse-support@ -- Simon Lees (Simotek) http://simotek.net Emergency Update Team keybase.io/simotek SUSE Linux Adelaide Australia, UTC+10:30 GPG Fingerprint: 5B87 DB9D 88DC F606 E489 CEC5 0922 C246 02F0 014B -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 21/11/2018 13.55, Stefan Seyfried wrote:

Am 21.11.18 um 13:33 schrieb Carlos E. R.:

...

Well, no, it seems systemd-timesyncd uses much more memory.

You are looking at the wrong field.

Nobody realy cares how much virtual memory a process has mapped / reserved, but instead how big its resident set size is.

I looked at the resident size :-) 76 chronyd 3024 systemd-timesyn

Example how to "use" 4G of memory:

seife@strolchi:/dev/shm> cat test.c #include #include #include #include #include

#define handle_error(msg) do { perror(msg); exit(EXIT_FAILURE); } while (0) int main(int argc, char *argv[]) { char *addr; size_t length = 1024L*1024*1024*4; /* 4GB */ int fd = open(argv[1], O_RDONLY); if (fd == -1) handle_error("open"); addr = mmap(NULL, length, PROT_READ, MAP_PRIVATE, fd, 0); if (addr == MAP_FAILED) handle_error("mmap");

while (1) sleep(1);

/* not reached */ munmap(addr, length); close(fd); return 0; } seife@strolchi:/dev/shm> gcc test.c -o test seife@strolchi:/dev/shm> truncate -s 10G test.file seife@strolchi:/dev/shm> ./test test.file & [1] 1905 seife@strolchi:/dev/shm> ps u 1905 USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND seife 1905 0.0 0.0 4198484 672 pts/10 S 13:52 0:00 ./test test.file

recompile with different "length=" and you can "use" even more.

Thanks for the example :-) -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)

On 21/11/2018 14.10, Stefan Seyfried wrote:

Am 21.11.18 um 14:00 schrieb Carlos E. R.:

...

On 21/11/2018 13.55, Stefan Seyfried wrote:

...

Am 21.11.18 um 13:33 schrieb Carlos E. R.:

...

Well, no, it seems systemd-timesyncd uses much more memory.

You are looking at the wrong field.

Nobody realy cares how much virtual memory a process has mapped / reserved, but instead how big its resident set size is.

I looked at the resident size :-)

Ok, I was confused by the "top" line.

...

76 chronyd 3024 systemd-timesyn

But something is fishy there. If even my pretty minimal example program uses 772kB (with 708kB shared) according to top,then I cannot imagine how chrony will work with just 76kb.

I don't know. But I guess it is waiting for an event to happen, then load the rest of the code. The "S" means it sleeping IIRC, so it matches. Ie, the clock is OK, and it is waiting some timer till checking the clock on internet again. Maybe after some hours running the systemd implementation does the same, dunno.

You are also using different tools on both machines (see the different output columns), so the results might not be comparable.

No, the testing was done on the same laptop, minutes of difference (you can see the uptime on the first line), and both times I used "top -b n1 | less". Once as user and the other as root by mistake, though. Why do you think it was different machines? :-?

It would of course be highly embarassing if a stupid sntp client like systemd-timesyncd really uses more memory than a full fledged ntp server implementation like chronyd. Judging from other systemd experiences, I would not be surprised, though.

:-} -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)

On 19/11/2018 14:09, Cristian Rodríguez wrote:

Let's say the traditional ntp daemon has not aged very well.. and is steadly bitrotting.

I believe it's a bit contentious, but is NTPsec relevant to this? -- Liam Proven - Technical Writer, SUSE Linux s.r.o. Corso II, Křižíkova 148/34, 186-00 Praha 8 - Karlín, Czechia Email: lproven@suse.com - Office telephone: +420 284 241 084 -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

* Marco Calistri [11-18-18 16:24]:

Il 18/11/18 16:56, Carlos E. R. ha scritto:

...

On 18/11/2018 14.51, Marco Calistri wrote:

...

...

I was relying on ntpd service since ever until I read about "chronyd" as the substitute of ntpd adopted by new opensuse releases, then I switched to chronyd.

Despite this it seems ntp package being still necessary (?) or could be removed by the system?

To verify if my system is synchronized I use the command "timedatectl":

marco@linux-turion64:~> timedatectl Local time: dom 2018-11-18 11:40:48 -02 Universal time: dom 2018-11-18 13:40:48 UTC RTC time: dom 2018-11-18 11:40:48 Time zone: America/Sao_Paulo (-02, -0200) System clock synchronized: yes NTP service: inactive RTC in local TZ: yes

Interesting command! I did not know about it. But it would be rather more interesting if it displayed milliseconds, as it does it is not sufficient.

It also, now that I think, does not display the network time at all, in my case from ntp, so I do not know how well my clock is synced - or not.

Hello Carlos,

My Linux box is not a 24/24 server then I decided to switch to chronyd: https://www.thegeekdiary.com/centos-rhel-7-chrony-vs-ntp-differences-between...

being a server is not a requirement for ntpd vs chronyd, it is a choice. one may choose either for any connected box providing any service. 18:07 toshiba: ~ # systemctl status chronyd;timedatectl;ntptime ● chronyd.service - NTP client/server Loaded: loaded (/usr/lib/systemd/system/chronyd.service; enabled; vendor preset: disabled) Active: inactive (dead) since Sun 2018-11-18 15:04:42 EST; 3h 3min ago Docs: man:chronyd(8) man:chrony.conf(5) Main PID: 2192 (code=exited, status=0/SUCCESS) Nov 18 11:31:34 toshiba.wahoo.no-ip.org systemd[1]: Starting NTP client/server... Nov 18 11:31:35 toshiba.wahoo.no-ip.org chronyd[2192]: chronyd version 3.4 starting (+CMDMON +NTP +REFCLOCK +RTC +PRIVD> Nov 18 11:31:35 toshiba.wahoo.no-ip.org chronyd[2192]: Frequency -15.548 +/- 0.076 ppm read from /var/lib/chrony/drift Nov 18 11:31:35 toshiba.wahoo.no-ip.org systemd[1]: Started NTP client/server. Nov 18 11:31:40 toshiba.wahoo.no-ip.org chronyd[2192]: Selected source 199.102.46.73 Nov 18 15:04:42 toshiba.wahoo.no-ip.org chronyd[2192]: chronyd exiting Nov 18 15:04:42 toshiba.wahoo.no-ip.org systemd[1]: Stopping NTP client/server... Nov 18 15:04:42 toshiba.wahoo.no-ip.org systemd[1]: Stopped NTP client/server. Local time: Sun 2018-11-18 18:08:30 EST Universal time: Sun 2018-11-18 23:08:30 UTC RTC time: Sun 2018-11-18 18:08:30 Time zone: America/Indiana/Indianapolis (EST, -0500) System clock synchronized: no NTP service: active RTC in local TZ: yes Warning: The system is configured to read the RTC time in the local time zone. This mode cannot be fully supported. It will create various problems with time zone changes and daylight saving time adjustments. The RTC time is never updated, it relies on external facilities to maintain it. If at all possible, use RTC in UTC by calling 'timedatectl set-local-rtc 0'. ntp_gettime() returns code 5 (ERROR) time df9c6f6e.4eed1394 Sun, Nov 18 2018 18:08:30.308, (.308305198), maximum error 1412500 us, estimated error 0 us, TAI offset 0 ntp_adjtime() returns code 5 (ERROR) modes 0x0 (), offset 12.379 us, frequency 18.604 ppm, interval 1 s, maximum error 1412500 us, estimated error 0 us, status 0x6041 (PLL,UNSYNC,NANO,MODE), time constant 7, precision 0.001 us, tolerance 500 ppm, 18:10 wahoo: ~ # systemctl status ntpd;timedatectl;ntptime ● ntpd.service - NTP Server Daemon Loaded: loaded (/usr/lib/systemd/system/ntpd.service; enabled; vendor preset: disabled) Drop-In: /run/systemd/generator/ntpd.service.d └─50-insserv.conf-$time.conf Active: active (running) since Sun 2018-11-18 18:05:22 EST; 5min ago Docs: man:ntpd(1) Process: 20973 ExecStart=/usr/sbin/start-ntpd start (code=exited, status=0/SUCCESS) Main PID: 20984 (ntpd) Tasks: 2 (limit: 512) CGroup: /system.slice/ntpd.service ├─20984 /usr/sbin/ntpd -p /var/run/ntp/ntpd.pid -g -u ntp:ntp -c /etc/ntp.conf └─20985 ntpd: asynchronous dns resolver Nov 18 18:05:22 wahoo systemd[1]: Starting NTP Server Daemon... Nov 18 18:05:22 wahoo ntpd[20983]: ntpd 4.2.8p10@1.3728-o Fri May 26 14:07:29 UTC 2017 (1): Starting Nov 18 18:05:22 wahoo ntpd[20983]: Command line: /usr/sbin/ntpd -p /var/run/ntp/ntpd.pid -g -u ntp:ntp -c /etc/ntp.conf Nov 18 18:05:22 wahoo ntpd[20984]: proto: precision = 0.084 usec (-23) Nov 18 18:05:22 wahoo ntpd[20984]: switching logging to file /var/log/ntp Nov 18 18:05:22 wahoo start-ntpd[20973]: Starting network time protocol daemon (NTPD) Nov 18 18:05:22 wahoo systemd[1]: Started NTP Server Daemon. Hint: Some lines were ellipsized, use -l to show in full. Local time: Sun 2018-11-18 18:11:15 EST Universal time: Sun 2018-11-18 23:11:15 UTC RTC time: Sun 2018-11-18 23:11:15 Time zone: America/Indiana/Indianapolis (EST, -0500) Network time on: yes NTP synchronized: yes RTC in local TZ: no ntp_gettime() returns code 0 (OK) time df9c7013.f22003ac Sun, Nov 18 2018 18:11:15.945, (.945801052), maximum error 169215 us, estimated error 1875 us, TAI offset 0 ntp_adjtime() returns code 0 (OK) modes 0x0 (), offset -836.944 us, frequency -8.556 ppm, interval 1 s, maximum error 169215 us, estimated error 1875 us, status 0x2001 (PLL,NANO), time constant 6, precision 0.001 us, tolerance 500 ppm, in this particular case the server happens to be running ntpd but could just as well be running chronyd. the time precision reported is the same. -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri Registered Linux User #207535 @ http://linuxcounter.net Photos: http://wahoo.no-ip.org/piwigo paka @ IRCnet freenode -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 19/11/2018 01.33, Marco Calistri wrote:

Il 18/11/18 21:14, Patrick Shanahan ha scritto:

...

in this particular case the server happens to be running ntpd but could just as well be running chronyd. the time precision reported is the same.

I personally dislike using two daemons doing practically the same job, if I can just use one.

I understand it is two computers: Nov 18 11:31:34 toshiba.wahoo.no-ip.org systemd[1]: Starting NTP client/server... Nov 18 11:31:35 toshiba.wahoo.no-ip.org chronyd[2192]: chronyd version 3.4 starting (+CMDMON +NTP +REFCLOCK +RTC +PRIVD> vs Nov 18 18:05:22 wahoo systemd[1]: Starting NTP Server Daemon... -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)

Il 19/11/18 01:18, Simon Lees ha scritto:

On 19/11/2018 00:21, Marco Calistri wrote:

...

Il 13/11/18 18:24, Achim Gratz ha scritto:

...

The handling of ntpd apparently switched from an old-style rc script to a systemd unit. Since then the stop script will take somewhere between ten seconds to a minute to declare success and continue with the shutdown. I don't see why it should take that long to shut down since ntpd really just closes the ntp logs and exits. The systemd unit file treats it as a forking daemon, which probably makes the check of whether the daemon has exited more involved. But ntpd can easily be told not to fork, so I think it should use the "-n" option in ExecStart and all the stuff that the current start script is doing should go into ExecStartPre?

Regards, Achim.

I was relying on ntpd service since ever until I read about "chronyd" as the substitute of ntpd adopted by new opensuse releases, then I switched to chronyd.

Despite this it seems ntp package being still necessary (?) or could be removed by the system?

It shouldn't be necessary into the future, there have been a couple of issues making that happen but hopefully it will be done soon if not now.

I had removed (zypper rm) ntpd some weeks ago but it went installed again at first zypper dup. Also would like to report a known *bug* or perhaps a not expected behaviour of chronyd about wrong permission warning on /var/run/chrony: sudo systemctl status chronyd ● chronyd.service - NTP client/server Loaded: loaded (/usr/lib/systemd/system/chronyd.service; enabled; vendor pre> Active: active (running) since Mon 2018-11-19 09:03:20 -02; 4min 32s ago Docs: man:chronyd(8) man:chrony.conf(5) Process: 1405 ExecStartPost=/usr/share/chrony-helper update-daemon (code=exit> Process: 1067 ExecStart=/usr/sbin/chronyd $OPTIONS (code=exited, status=0/SUC> Main PID: 1103 (chronyd) Tasks: 1 (limit: 4592) Memory: 1.2M CGroup: /system.slice/chronyd.service └─1103 /usr/sbin/chronyd nov 19 09:03:12 linux-turion64 systemd[1]: Starting NTP client/server... nov 19 09:03:13 linux-turion64 chronyd[1103]: chronyd version 3.4 starting (+CM> nov 19 09:03:13 linux-turion64 chronyd[1103]: Wrong permissions on /var/run/chr> nov 19 09:03:13 linux-turion64 chronyd[1103]: Disabled command socket /var/run/> nov 19 09:03:13 linux-turion64 chronyd[1103]: Frequency -12.652 +/- 26.222 ppm > nov 19 09:03:20 linux-turion64 systemd[1]: Started NTP client/server. It has been reported also for other distros, I mean it is not a specific opensuse's chrony error. In addition I also would like to ask about the second warning: "Disabled command socket /var/run/" is it an error we can correct in some way or it is just an informative message? Thanks and regards, -- Marco Calistri Build: openSUSE Tumbleweed 20181116 Kernel: 4.19.1-1-default - Cinnamon 3.8.9