hardy Abrupt clock drift with high CPU usage

Bug #185632 reported by Rodrigo Azevedo
This bug report is a duplicate of:  Bug #17589: System clock runs far too fast. Edit Remove
6
Affects Status Importance Assigned to Milestone
linux (Ubuntu)
Incomplete
Undecided
Unassigned

Bug Description

Binary package hint: linux-image-2.6.24-4-server

When I upgrade from 7.10 to hardy my clock simple drift few seconds/min.
With high cpu usage the clock drift 5min or 10min at random!
The problem is consistent with 4 machines (same hardware) I upgraded.

The NTP can't sync the clock and ntpdate running every min drift tons of seconds.

The unique warning in logs is the infamous "Clocksource tsc unstable (delta = 732182695 ns)" 5min later the system's boot.

I try "noapic" and "notsc" but the problem persist.

AMD Athlon(tm) 64 X2 Dual Core Processor 4600+
2.6.24-4-server #1 SMP Mon Jan 14 18:47:45 UTC 2008 x86_64 GNU/Linux

When a big change happen in the clock system "freezes" and only returns to work properly when the cpu diminish their use.

Rodrigo

Revision history for this message
Rodrigo Azevedo (rodrigoams) wrote :
Download full text (6.4 KiB)

With the options "noapic" and "notsc" plus ntp-server the clock remains sync but with high cpu usage... look this:

  JOBID PARTITION NAME USER ST TIME NODES NODELIST(REASON)
    361 pnc Bet7_pon pnc R 1:08:07 4 teoria,teoria[1-3]
    360 testes teste1MA rodrigo R 1:59:50 2 teoria[5,8]
teoria9: Qui Jan 24 21:36:54 BRT 2008
teoria1: Qui Jan 24 21:36:54 BRT 2008
teoria2: Qui Jan 24 21:36:54 BRT 2008
teoria5: Qui Jan 24 21:36:54 BRT 2008
teoria3: Qui Jan 24 21:36:54 BRT 2008
teoria7: Qui Jan 24 21:36:54 BRT 2008
teoria4: Qui Jan 24 21:36:54 BRT 2008

-> teoria8: Qui Jan 24 20:26:36 BRT 2008 <-

teoria10: Qui Jan 24 21:36:55 BRT 2008
teoria6: Qui Jan 24 21:36:55 BRT 2008

After 2:00 hs the teoria8's clock delayed 1:10 hs!

rodrigo@teoria8:~$ ps aux
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
root 1 0.0 0.0 4020 864 ? Ss 10:00 0:00 /sbin/init
root 2 0.0 0.0 0 0 ? S< 10:00 0:00 [kthreadd]
root 3 0.0 0.0 0 0 ? S< 10:00 0:00 [migration/0]
root 4 0.0 0.0 0 0 ? S< 10:00 0:00 [ksoftirqd/0]
root 5 0.0 0.0 0 0 ? S< 10:00 0:00 [watchdog/0]
root 6 0.0 0.0 0 0 ? S< 10:00 0:00 [migration/1]
root 7 0.0 0.0 0 0 ? S< 10:00 0:00 [ksoftirqd/1]
root 8 0.0 0.0 0 0 ? S< 10:00 0:00 [watchdog/1]
root 9 0.0 0.0 0 0 ? S< 10:00 0:00 [events/0]
root 10 0.0 0.0 0 0 ? S< 10:00 0:00 [events/1]
root 11 0.0 0.0 0 0 ? S< 10:00 0:00 [khelper]
root 45 0.0 0.0 0 0 ? S< 10:00 0:00 [kblockd/0]
root 46 0.0 0.0 0 0 ? S< 10:00 0:00 [kblockd/1]
root 49 0.0 0.0 0 0 ? S< 10:00 0:00 [kacpid]
root 50 0.0 0.0 0 0 ? S< 10:00 0:00 [kacpi_notify]
root 184 0.0 0.0 0 0 ? S< 10:00 0:00 [kseriod]
root 233 0.0 0.0 0 0 ? S 10:00 0:00 [pdflush]
root 234 0.0 0.0 0 0 ? S 10:00 0:00 [pdflush]
root 235 0.0 0.0 0 0 ? S< 10:00 0:00 [kswapd0]
root 279 0.0 0.0 0 0 ? S< 10:00 0:00 [aio/0]
root 280 0.0 0.0 0 0 ? S< 10:00 0:00 [aio/1]
root 1491 0.0 0.0 0 0 ? S< 10:00 0:00 [ksuspend_usbd]
root 1494 0.0 0.0 0 0 ? S< 10:00 0:00 [khubd]
root 1512 0.0 0.0 0 0 ? S< 10:00 0:00 [ata/0]
root 1515 0.0 0.0 0 0 ? S< 10:00 0:00 [ata/1]
root 1518 0.0 0.0 0 0 ? S< 10:00 0:00 [ata_aux]
root 1587 0.0 0.0 0 0 ? S< 10:00 0:00 [khpsbpkt]
root 1863 0.0 0.0 0 0 ? S< 10:00 0:00 [scsi_eh_0]
root 1864 0.0 0.0 0 0 ? S< 10:00 0:00 [scsi_eh_1]
root 2324 0.0 0.0 0 0 ? S< 10:00 0:00 [scsi_eh_2]
root 2325 0.0 0.0 0 ...

Read more...

Revision history for this message
Leann Ogasawara (leannogasawara) wrote :

Thank you for taking the time to report this bug and helping to make Ubuntu better. Per the kernel team's bug policy, can you please attach the following information. Please be sure to attach each file as a separate attachment.

* uname -a > uname-a.log
* cat /proc/version_signature > version.log
* dmesg > dmesg.log
* sudo lspci -vvnn > lspci-vvnn.log

For more information regarding the kernel team bug policy, please refer to https://wiki.ubuntu.com/KernelTeamBugPolicies . Thanks again and we appreciate your help and feedback.

Changed in linux:
status: New → Incomplete
Revision history for this message
Rodrigo Azevedo (rodrigoams) wrote :
Revision history for this message
Rodrigo Azevedo (rodrigoams) wrote :
Revision history for this message
Rodrigo Azevedo (rodrigoams) wrote :
Revision history for this message
Rodrigo Azevedo (rodrigoams) wrote :
Revision history for this message
Rodrigo Azevedo (rodrigoams) wrote :

With "acpi=off noapic nolapic" boot options the problem dissapears. The clock remain sync, the network interface stay stable (other problem/bug) and the computer no more "freeze".

Other idea?

Rodrigo

Revision history for this message
Shivcharan (nashikit) wrote :

Hi,
I am new in linux and I have also problem of "tsc clock unstable ". Please help me for solving this problem;
How I can edit "acpi=off noapic nolapic"?

Revision history for this message
Leann Ogasawara (leannogasawara) wrote :

Hi Rodrigo,

Can you also attach your dmesg output when not booting with those options? Thanks.

Schivchiran, in order to edit your boot options: "Press 'Escape' key at the 3 second pause by Grub bootloader. Then press 'e' (edit) on the buggy kernel entry, followed by 'e' again on the kernel line. Then remove the splash and quiet keywords and press 'b' to boot."

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.