Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
2.6.10 breaks ntpd?
View unanswered posts
View posts from last 24 hours

Goto page 1, 2  Next  
Reply to topic    Gentoo Forums Forum Index Kernel & Hardware
View previous topic :: View next topic  
Author Message
energy
n00b
n00b


Joined: 07 Dec 2004
Posts: 25
Location: Oulu, Finland

PostPosted: Wed Jan 19, 2005 6:15 pm    Post subject: 2.6.10 breaks ntpd? Reply with quote

It seems that using kernel 2.6.10 breaks the ntpd. I've tried stock 2.6.10, 2.6.10-cko3 and 2.6.10-morph12. NTPD just won't keep syncing the clock, there are no error messages in any log files... This is very strange since I've used the same kernel config with 2.6.9-XXX and it works perfectly!

Any ideas? I've tried almost everything... anyone having the same problem? :roll:
Back to top
View user's profile Send private message
pilla
Bodhisattva
Bodhisattva


Joined: 07 Aug 2002
Posts: 7729
Location: Underworld

PostPosted: Wed Jan 19, 2005 8:40 pm    Post subject: Reply with quote

I am not using ntpd, but my clock went to 7 o'clock sometimes. Just too weird.

I had to give up 2.6.10/11 because of problems with suspend and usb mouse too.
_________________
"I'm just very selective about the reality I choose to accept." -- Calvin
Back to top
View user's profile Send private message
energy
n00b
n00b


Joined: 07 Dec 2004
Posts: 25
Location: Oulu, Finland

PostPosted: Thu Jan 20, 2005 2:18 pm    Post subject: Reply with quote

The ntpd is not the only problem, the clock starts to run "overspeed" e.g. after a day it is at least +20 minutes compared to real time. This problem came bundled with the ntpd problem...
Back to top
View user's profile Send private message
toralf
Developer
Developer


Joined: 01 Feb 2004
Posts: 3922
Location: Hamburg

PostPosted: Thu Jan 20, 2005 3:25 pm    Post subject: Reply with quote

I did not have any problems using the following kernels: linux-2.6.10-gentoo-r6 linux-2.6.10-hardened-r1 linux-2.6.10-r1
But be sure you did not have a line in your ntp.conf like:
restrict default ignore
Back to top
View user's profile Send private message
drescherjm
Advocate
Advocate


Joined: 05 Jun 2004
Posts: 2790
Location: Pittsburgh, PA, USA

PostPosted: Thu Jan 20, 2005 8:23 pm    Post subject: Reply with quote

Quote:
The ntpd is not the only problem, the clock starts to run "overspeed" e.g. after a day it is at least +20 minutes compared to real time


I have the same problem. It drives me absolutely nuts. However I believe it existed in previous kernels < 2.6.10. In a period of a week (not connected to the internet) my clock can be off by as much as + or - 3 hours. At one point I thought it was rebooting and timezones but then this does not explain how the clock was ahead by 87 minutes... My rig is a dual processor athlon MX 2200 with 2 GB of PC2100 ECC DDR. Now running SMP gentoo-dev-sources 2.6.10r2.
_________________
John

My gentoo overlay
Instructons for overlay
Back to top
View user's profile Send private message
markandrew
Apprentice
Apprentice


Joined: 04 Feb 2004
Posts: 176
Location: Manchester, UK

PostPosted: Fri Jan 21, 2005 1:40 am    Post subject: Reply with quote

two things i noticed about ntp recently:

1) i have 2 NICs, eth0 and ath0; ntp requires both of them to be active to start. apparently this will be be fixed in the next baselayout

2) i also could not sync today whatever i tried. i noticed that my /etc/ntp.conf file was being overwritten everytime i ran ntp, removing my server list. so i copied it from the one in /usr/share, added more servers, and removed all write perms on it. still didn't work. in desperation i disabled my ath0 NIC by removing the symlink net.ath0 -> net.lo in /etc/init.d/ ... now it works. strange huh?
Back to top
View user's profile Send private message
JjcampNR
Tux's lil' helper
Tux's lil' helper


Joined: 02 Mar 2003
Posts: 127
Location: Massachusetts, USA

PostPosted: Fri Jan 21, 2005 7:01 am    Post subject: Reply with quote

I've also been having problems with a jumping clock, it's so damn annoying. I'm running 2.6.10-gentoo-r3. I'm going to try moving to r6 and see if that fixes anything.

--Josh
_________________
Damn the man, fight the power!

http://www.torquenetworks.com
Back to top
View user's profile Send private message
energy
n00b
n00b


Joined: 07 Dec 2004
Posts: 25
Location: Oulu, Finland

PostPosted: Mon Jan 24, 2005 8:18 pm    Post subject: Reply with quote

Anyone had luck finding the cause of the problem?

I wonder if it's a bug with certain chipsets or just an option with kernel that doesn't work in combination with other option(s)... I have a nForce chipset.
Back to top
View user's profile Send private message
pilla
Bodhisattva
Bodhisattva


Joined: 07 Aug 2002
Posts: 7729
Location: Underworld

PostPosted: Mon Jan 24, 2005 9:04 pm    Post subject: Reply with quote

has anybody tried 2.6.11-rc2?
_________________
"I'm just very selective about the reality I choose to accept." -- Calvin
Back to top
View user's profile Send private message
drescherjm
Advocate
Advocate


Joined: 05 Jun 2004
Posts: 2790
Location: Pittsburgh, PA, USA

PostPosted: Tue Jan 25, 2005 4:15 am    Post subject: Reply with quote

I have the original AMD MP chipset on my dual processor TYAN Thunder AMD K7 2200 MP rig.
_________________
John

My gentoo overlay
Instructons for overlay
Back to top
View user's profile Send private message
AlterEgo
Veteran
Veteran


Joined: 25 Apr 2002
Posts: 1619

PostPosted: Tue Jan 25, 2005 8:33 pm    Post subject: Reply with quote

Try openntpd.
It's easier, and it behaves 8) on any kernel I've tried.
Back to top
View user's profile Send private message
energy
n00b
n00b


Joined: 07 Dec 2004
Posts: 25
Location: Oulu, Finland

PostPosted: Wed Feb 09, 2005 1:54 pm    Post subject: Reply with quote

Tried openntpd, tried about everything... for a moment I thought I found a solution but no.

If sync time manually e.g. "ntpdate tick.keso.fi" it works and clock is in sync. The ntpd itself finds the servers and keeps track of offset but just won't keep syncing. I used debug and log files reveal these errors for every server:

Code:
 9 Feb 15:48:01 ntpd[9195]: peer 130.233.224.2 event 'event_reach' (0x84) status 'unreach, conf, 1 event, event_reach' (0xa014)


As I said it worked with earlier kernel version so there is nothing wrong with my configs. I've tried different configurations with no success.. here's my current ntp.conf if someone wants to take a look:

Code:

# Add some useful log messages to our ntpd.log file
#
logconfig       =syncevents +peerevents +sysevents +allclock
logfile         /var/log/ntpd.log
driftfile       /var/lib/ntp/ntp.drift

# Allow local clients to query this server, but not update it
#
restrict default ignore
restrict 127.0.0.1
restrict 192.168.1.0 mask 255.255.255.0 nomodify nopeer notrap

# Stratum-2 servers
#
server tock.keso.fi
restrict tock.keso.fi nomodify notrap noquery


This one is taken from some thread in this forum.. my original config was a bit different but it didn't matter. I've tried googling about this problem but no-one seems to know what's causing it.

EDIT: Removed all but one server from the list
Back to top
View user's profile Send private message
didl
Retired Dev
Retired Dev


Joined: 09 Sep 2003
Posts: 1106
Location: Pittsburgh, PA

PostPosted: Wed Feb 09, 2005 3:14 pm    Post subject: Reply with quote

I also had trouble with ntpd using 2.6.10. However, it turned out that
the problem was ntpd droping privileges from user root to ntp. I
was able to fix this via emerging ntpd with
Code:

USE=+nodroproot


HTH
Back to top
View user's profile Send private message
sl70
Guru
Guru


Joined: 18 Jun 2002
Posts: 449
Location: Saitama, JP

PostPosted: Sat Feb 12, 2005 5:29 pm    Post subject: Reply with quote

I've also been having the same problem since about a month or so ago. I have no idea why it should have started all of a sudden.

Last week I uninstalled ntp and installed openntp. This seemed to work for a while but yesterday I started getting these messages in my log:

Code:
Feb 11 12:19:32 musume ntpd[31832]: peer 199.184.165.135 now valid
Feb 11 12:20:02 musume ntpd[31832]: peer 199.184.165.135 now invalid
Feb 11 12:20:17 musume ntpd[31832]: peer 199.184.165.135 now valid
Feb 11 12:21:21 musume ntpd[31831]: adjusting local clock by -2144.407760s
Feb 11 12:24:53 musume ntpd[31831]: adjusting local clock by -2145.419356s
Feb 11 12:28:44 musume ntpd[31831]: adjusting local clock by -2146.442124s
Feb 11 12:28:44 musume ntpd[31831]: adjtime failed: Invalid argument
Feb 11 12:32:21 musume ntpd[31831]: adjusting local clock by -2147.515935s
Feb 11 12:32:21 musume ntpd[31831]: adjtime failed: Invalid argument


Since that time whenever ntpd tries to adjust the time it puts out an invalid argument message. And now after about a day my clock is 45 minutes fast. Here's my ntpd.conf:

Code:
# Addresses to listen on (ntpd does not listen by default)

listen on 192.168.1.0

# sync to a single server
#server ntp.example.org

# use a random selection of 8 public stratum 2 servers
# see http://twiki.ntp.org/bin/view/Servers/NTPPoolServers
servers pool.ntp.org


What is going on??? :?
Back to top
View user's profile Send private message
sl70
Guru
Guru


Joined: 18 Jun 2002
Posts: 449
Location: Saitama, JP

PostPosted: Sun Feb 13, 2005 1:53 am    Post subject: Reply with quote

I upgraded the kernel to 2.6.10-r6 and now the clock is running right. I had a little adjustment of 0.7 sec when I booted up and ntpd started, but nothing since then.

Whew! Glad that's fixed.
Back to top
View user's profile Send private message
energy
n00b
n00b


Joined: 07 Dec 2004
Posts: 25
Location: Oulu, Finland

PostPosted: Sun Feb 13, 2005 9:03 am    Post subject: Reply with quote

I tried 2.6.10-r6 and it didn't help, ntpd didn't work neither did the clock.
However I think I found out why.. there seems to be problem with nForce-
chipset and timer. With "working" kernel there are these messages at boot:

Code:

Feb 11 23:19:47 chevron ENABLING IO-APIC IRQs
Feb 11 23:19:47 chevron ..TIMER: vector=0x31 pin1=2 pin2=-1
Feb 11 23:19:47 chevron ..MP-BIOS bug: 8254 timer not connected to IO-APIC
Feb 11 23:19:47 chevron ...trying to set up timer (IRQ0) through the 8259A ...  failed.
Feb 11 23:19:47 chevron ...trying to set up timer as Virtual Wire IRQ... failed.
Feb 11 23:19:47 chevron ...trying to set up timer as ExtINT IRQ... works.


This part is missing completely with .10 kernels... also I took a look at
.9 -> .10 changelog at kernel.org. There has been a lot of forking with timer code.
Back to top
View user's profile Send private message
sl70
Guru
Guru


Joined: 18 Jun 2002
Posts: 449
Location: Saitama, JP

PostPosted: Sun Feb 13, 2005 9:59 pm    Post subject: Reply with quote

Ack. Posted too soon. I'm not getting those ntp messages anymore, but my clock is still running too fast. I guess all we can do is wait until a new kernel comes out.
Back to top
View user's profile Send private message
pompafrittes
n00b
n00b


Joined: 29 Feb 2004
Posts: 39
Location: Sweden

PostPosted: Wed Mar 02, 2005 7:12 pm    Post subject: Reply with quote

sl70 wrote:
Ack. Posted too soon. I'm not getting those ntp messages anymore, but my clock is still running too fast. I guess all we can do is wait until a new kernel comes out.


Got those messages too before but they appeared because the time was one hour wrong. when i set the time a little closer to correct time i got thiese instead

Code:

Mar  2 19:55:55 src@sture ntpd[29373]: adjusting local clock by 260.723623s
Mar  2 19:58:26 src@sture ntpd[29373]: adjusting local clock by 260.604434s
Mar  2 20:01:24 src@sture ntpd[29373]: adjusting local clock by 260.319579s
Mar  2 20:05:25 src@sture ntpd[29373]: adjusting local clock by 260.169114s


but they will dissapear when the time is correct.
_________________
Doh
Back to top
View user's profile Send private message
gtbX
Tux's lil' helper
Tux's lil' helper


Joined: 11 Oct 2003
Posts: 126

PostPosted: Sat Mar 05, 2005 6:45 pm    Post subject: Reply with quote

I've been having this problem with all 3 of my gentoo boxes. The NForce1 box will run an hour fast after a week, while the PII fileserver will be an hour slow. My laptop is even stranger - after I bring it out of sleep mode, the clock is ~30 hours faster than it should be. All are running 2.6.10-r6 (except for the laptop, only going back to 2.6.9-r13 would fix it).

I was going to install ntpd as a stopgap, but this thread is making me think twice.
Back to top
View user's profile Send private message
Legoguy
Apprentice
Apprentice


Joined: 22 Dec 2003
Posts: 166
Location: Edmonton, Alberta, Canada

PostPosted: Sat Mar 05, 2005 9:33 pm    Post subject: Reply with quote

Try enabling the Power Management Timer Support in ACPI for the laptop.

Everyone else + the laptop: Try enabling Enhanced Real Time Clock in Device Drivers -> Character Devices. If it's already checked... then, well.. I can't really think what might be the cause of this (I don't experience it..)
Back to top
View user's profile Send private message
pilla
Bodhisattva
Bodhisattva


Joined: 07 Aug 2002
Posts: 7729
Location: Underworld

PostPosted: Mon Mar 07, 2005 6:28 pm    Post subject: Reply with quote

has anybody tested 2.6.11 final?
_________________
"I'm just very selective about the reality I choose to accept." -- Calvin
Back to top
View user's profile Send private message
timmy
n00b
n00b


Joined: 24 May 2002
Posts: 56
Location: Bath, UK

PostPosted: Tue Mar 08, 2005 12:38 pm    Post subject: Reply with quote

pilla wrote:
has anybody tested 2.6.11 final?

Yes, and I'm still seeing this problem with both ntp and openntpd on my nforce2 / Athlon 2500+ machine.

I'm just waiting for 2005.0 before installing on my new Dell Dimension 5000 Pentium4. It'll be interesting to see if it also displays the same problem.

Tim
Back to top
View user's profile Send private message
timmy
n00b
n00b


Joined: 24 May 2002
Posts: 56
Location: Bath, UK

PostPosted: Thu Mar 10, 2005 9:06 am    Post subject: Reply with quote

I've managed to get ntp working on my nforce2 machine: It seems my clock was running just too fast for ntp or openntpd to work.

The answer was to use /usr/bin/tickadj to reduce the tick (to 9935 in my case) as described here.

There's a post on gentoo forums here that has a patch for /etc/init.d/ntpd that will automatically apply and save the adjusted tick.

Tim
Back to top
View user's profile Send private message
energy
n00b
n00b


Joined: 07 Dec 2004
Posts: 25
Location: Oulu, Finland

PostPosted: Thu Mar 10, 2005 12:55 pm    Post subject: Reply with quote

Finally! Thanks to timmy I got it working using using tickadj.
Right value for me seems to be 9960.

But I really would like to know why this problem appears only
when moving to higher kernel version than 2.6.9? What did the
developers break this time? :)
Back to top
View user's profile Send private message
kallamej
Administrator
Administrator


Joined: 27 Jun 2003
Posts: 4975
Location: Gothenburg, Sweden

PostPosted: Fri Mar 11, 2005 9:42 pm    Post subject: Reply with quote

One of the problems with newer kernels is explained here. Tried to switch to openntpd, but it seems to just pick a random time stamp to send to the servers and doesn't sync either as described here, for instance. Back to 2.6.7...

Edit: The random time stamps is by design.
_________________
Please read our FAQ Forum, it answers many of your questions.
irc: #gentoo-forums on irc.libera.chat


Last edited by kallamej on Sat Mar 12, 2005 8:32 am; edited 1 time in total
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Kernel & Hardware All times are GMT
Goto page 1, 2  Next
Page 1 of 2

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum