Re: Clock Problem?

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

Unruh wrote:
gaedhealic <gaedhealic@xxxxxxxxx> writes:

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

Unruh wrote:
gaedhealic <gaedhealic@xxxxxxxxx> writes:

Now, using NTP, on boot the clock is set to the correct time, but again,
it runs fast.
That's a bit strange, but I've never used NTP. It looks as if
on shutdown the time was read by NTP from a time server.

Someone with more NTP experience than me should tell us about that.
agreed.
but it seems that time is set by NTP at bootup -- or when you select and
active NTP server -- then saved to the system on shutdown
With a system which runs at double speed, ntp will make no difference. I
have no idea how you are running ntp. If you are doing
ntpdate -b
one minute or something, then each time ntp will bring the clock back into
sync. On shutdown the on board system clock is copied to the hardware
clock.

If you are running ntpd then it should do nothing because 2000000 ppm
cannot be corrected by ntp-- either by frequency drift or by tick
rescaling.


doing ntpdate -b from a konsole command line does correct the display clock.

also, doing yast > network services > NTP configuration > During boot,
use random, finish, also corrects the display time

(this event, display time was 65 minutes fast)

But it does not keep the time correct. I mean, it does it
(corrects/resets the time) once, then the system display time runs fast.

Yes. ntpdate -b does a hard reset of the time. It simply resets the system
time to current ntp time. It does nothing to the rate or anything else.

ntp cannot correct the rate error in a clock that is running at double
speed.

One more suggestion.

cat /proc/interrupts
wait 10 sec according to your watch ( as accurately as you can) and then do
cat /proc/interrupts
again. (Pretype the command and hit the enter key on that 10 sec time )
Look at the first line, the timer interrupts. There should have been
2500 extra interrupts in that 10 sec. (Obviously your reaction time is not
perfect so it will probably be somewhere between 2400 and 2600 extra
interrupts.) If it is around 5000 extra then we know what the problem is.






first attempt---

#
Matchless:/home/owd # cat /proc/interrupts
CPU0
0: 1442001 IO-APIC-edge timer
1: 6849 IO-APIC-edge i8042
2: 0 XT-PIC cascade
6: 5 IO-APIC-edge floppy
8: 0 IO-APIC-edge rtc
12: 185258 IO-APIC-edge i8042
14: 66567 IO-APIC-edge ide0
15: 50310 IO-APIC-edge ide1
129: 13361 IO-APIC-level ehci_hcd:usb1, ohci_hcd:usb2, ohci_hcd:usb3
137: 764 IO-APIC-level eth0
145: 7804 IO-APIC-level ohci1394, snd_ca0106
153: 0 IO-APIC-level libata
161: 269 IO-APIC-level libata
NMI: 492
LOC: 720951
ERR: 2
MIS: 0

# cat /proc/interrupts
CPU0
0: 1446623 IO-APIC-edge timer
1: 6859 IO-APIC-edge i8042
2: 0 XT-PIC cascade
6: 5 IO-APIC-edge floppy
8: 0 IO-APIC-edge rtc
12: 185258 IO-APIC-edge i8042
14: 66582 IO-APIC-edge ide0
15: 50472 IO-APIC-edge ide1
129: 13397 IO-APIC-level ehci_hcd:usb1, ohci_hcd:usb2, ohci_hcd:usb3
137: 764 IO-APIC-level eth0
145: 7804 IO-APIC-level ohci1394, snd_ca0106
153: 0 IO-APIC-level libata
161: 269 IO-APIC-level libata
NMI: 492
LOC: 723262
ERR: 2
MIS: 0

1446623 - 1442001 = 4622

(I may have been a little under 10 seconds)

Second attempt --

# cat /proc/interrupts
CPU0
0: 1599121 IO-APIC-edge timer
1: 7026 IO-APIC-edge i8042
2: 0 XT-PIC cascade
6: 5 IO-APIC-edge floppy
8: 0 IO-APIC-edge rtc
12: 206907 IO-APIC-edge i8042
14: 67096 IO-APIC-edge ide0
15: 55836 IO-APIC-edge ide1
129: 14616 IO-APIC-level ehci_hcd:usb1, ohci_hcd:usb2, ohci_hcd:usb3
137: 852 IO-APIC-level eth0
145: 8080 IO-APIC-level ohci1394, snd_ca0106
153: 0 IO-APIC-level libata
161: 269 IO-APIC-level libata
NMI: 503
LOC: 799512
ERR: 2
MIS: 0
Matchless:/home/owd # cat /proc/interrupts
CPU0
0: 1604299 IO-APIC-edge timer
1: 7036 IO-APIC-edge i8042
2: 0 XT-PIC cascade
6: 5 IO-APIC-edge floppy
8: 0 IO-APIC-edge rtc
12: 206907 IO-APIC-edge i8042
14: 67154 IO-APIC-edge ide0
15: 56034 IO-APIC-edge ide1
129: 14656 IO-APIC-level ehci_hcd:usb1, ohci_hcd:usb2, ohci_hcd:usb3
137: 852 IO-APIC-level eth0
145: 8080 IO-APIC-level ohci1394, snd_ca0106
153: 0 IO-APIC-level libata
161: 269 IO-APIC-level libata
NMI: 504
LOC: 802101
ERR: 2
MIS: 0
Matchless:/home/owd #

1604299 - 1599121 = 5178
(a little over ten seconds?)

g

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)
Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org

iD8DBQFGL90xAqk3z+s2AjQRAt2NAJ0RXxXc+C4dYHKLNOBub5EpbqCqdwCfaPW7
XONCoM7kpyTfR21lvd8B4IU=
=JEE+
-----END PGP SIGNATURE-----
.

Relevant Pages

  • Re: Installing modem
    ... 0: 50953 IO-APIC-edge timer ... 12: 19237 IO-APIC-edge PS/2 Mouse ... VIA Technologies, Inc. UHCI USB ... d400-d41f: usb-uhci ...
    (Debian-User)
  • RE: Slow Network Connection
    ... Something I found interesting...PCLinuxOS out of the ones listed below is the only one that worked for me (normal network speeds). ... 6: 1 1 IO-APIC-edge floppy ... 0: 84 1 IO-APIC-edge timer ... 21: 0 0 IO-APIC-fasteoi libata ...
    (Debian-User)
  • PCI Interrupt
    ... I understand that PCI interrupts should be level rather than edge ... 0: 1614601 IO-APIC-edge timer ... someone other than recipient is not authorized and may be unlawful. ...
    (Linux-Kernel)
  • Re: SATA RAID stoert DVB
    ... Besserung zu bringen scheint. ... 0: 102739 IO-APIC-edge timer ... 15: 3517 IO-APIC-edge ide1 ...
    (de.comp.os.unix.linux.hardware)