Last week I tried bogus values in the drift file: no luck (didn't try zero,
though). Now I tried using your drift file: still no luck. Then I tried
with your chrony.conf as well (with the server changed to mine). Bingo!
Looks like a zero error bound plus "initstepslew 02" is required to trigger
t
> Interestingly, though, I'm not sure how that would cause the problem
> myself.
The zero error bound, after a long chain of events that I will have to
trace out, causes an NaN (invalid float) to be assigned to "skew". This
propagates to adjtimex() along with an invalid "tick" value. I thought
t
> What do you have in /var/lib/chrony/chrony.drift? Have you tried deleting
> that file?
Interesting. The file is actually /etc/chrony.drift on my system (Chrony on
this computer dates back a long time, possibly 8 years, and I've just not got
around to moving it). However, removing this file work
What do you have in /var/lib/chrony/chrony.drift? Have you tried deleting
that file?
--
John Hasler
j...@dhh.gt.org
Elmwood, WI USA
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
> Could you run 'sudo chrony -d' and send me any messages it prints before it
> crashes?
Certainly. I think they'll be in with the strace outputs, but difficult the
filter out from the output of strace. Output from 1.23-6 attached as requested.
All the best,
-Dave.
sys_linux.c:649:(get_version_sp
Could you run 'sudo chrony -d' and send me any messages it prints before it
crashes?
--
John Hasler
j...@dhh.gt.org
Elmwood, WI USA
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Sorry for the delay in responding today. Since I've got a few things to do,
I'll put the responses together here. Unfortunately, I won't be around over
the next couple of days, but if you have any requests I'll get on to them as
soon as I can next week.
> 1) Where did you get your 2.6.27 kernel?
What is the value of HZ in your kernel?
--
John Hasler
j...@dhh.gt.org
Elmwood, WI USA
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
It appears that TMX_SetFrequency() (and therefor adjtimex()) is being
called with an invalid value for "tick".
--
John Hasler
j...@dhh.gt.org
Elmwood, WI USA
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@list
> Just as a bit of additional information, it appears that chronyd will
> randomly die after starting without a connection from chronyc, but this
> isn't easy to reproduce. The method I've already described of executing
> "burst 5/10" remains the most reliable way to reproduce this problem on
> my
> It appears that adjtimex() is being called with garbage instead of a valid
> pointer to a timex structure.
Just as a bit of additional information, it appears that chronyd will
randomly die after starting without a connection from chronyc, but this
isn't easy to reproduce. The method I've alread
It appears that adjtimex() is being called with garbage instead of a valid
pointer to a timex structure. This probably means that something is
scribbling on the stack.
--
John Hasler
j...@dhh.gt.org
Elmwood, WI USA
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a
Thank you. This trace will be quite useful.
--
John Hasler
j...@dhh.gt.org
Elmwood, WI USA
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
>> If there is anything else you'd like me to do, feel free to get in touch
>> and I'll try thing out as soon as possible.
> Please run
>strace chronyd -d
> And send me the output.
Certainly. Please find the results log of this attached to this email. I'm
afraid I'm not experienced enough
> Not a problem at all. I redirected standard output and standard error to
> a file, and have attached it to this message. Chronyd segfaulted at
> exactly the same point (when "burst" was issued to chronyc).
And produced interesting and useful but unexpected results.
> If there is anything else y
> Please see if you can reproduce the bug running chronyd under strace like
> this:
> strace -eioctl chronyd -d
Not a problem at all. I redirected standard output and standard error to a
file, and have attached it to this message. Chronyd segfaulted at exactly the
same point (when "burst" was iss
Please see if you can reproduce the bug running chronyd under strace like
this:
strace -eioctl chronyd -d
And send me the output.
--
John Hasler
j...@dhh.gt.org
Elmwood, WI USA
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Co
Package: chrony
Version: 1.23-6
Severity: important
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA512
Hello,
I've just noticed that one of my systems had its clock out of sync with
that of the timeserver. The timeserver is another Debian machine running
chrony (also version 1.23-6). On investig
18 matches
Mail list logo
