Re: [chrony-dev] What is actually being done in "make check" for 113-leapsecond?

[ Thread Index | Date Index | More chrony.tuxfamily.org/chrony-dev Archives ]


On Tue, Mar 22, 2016 at 04:42:47PM -0600, Mark Post wrote:
> At SUSE, we use the Open Build Service (OBS) to build and package our software.  I'm experiencing an issue with the OBS during "make check" processing.  The 113-leapsecond test fails, which does not happen when I build it elsewhere.  To be able to explain to our OBS support team just what's going on so we can investigate this problem, could someone tell me what all is being done during that test?  The following is the output I get when the failure happens:

This test checks if chronyd inserts the leap second correctly using
the system driver. It looks like it didn't insert it at all as there
is one second error reported on the BAD line. Maybe chronyd didn't
know there was supposed to be a leap second, because the right/UTC
timezone was not installed on the build system? Is it in the tzdata
package?

As a quick check, you can run
$ TZ=right/UTC date -d 'Dec 31 2008 23:59:60'

If it doesn't print "Wed Dec 31 23:59:60 UTC 2008", there is a problem
with the timezone.

> 113-leapsecond (22/31)
> Testing leap second:
>   network with 1*1 servers and 1 clients:
>     non-default settings:
>       client_conf=leapsecmode system
>       limit=345600
>       max_sync_time=172800
>       refclock_jitter=1e-9
>       refclock_offset=(* -1.0 (equal 0.1 (max (sum 1.0) 172800) 172800))
>       server_conf=refclock SHM 0 dpoll 10 poll 10 leapsectz right/UTC

>     checking clock sync time, max/rms time/freq error:
>       node 2: 174364 1.00e+00 5.09e-02 8.27e-02 4.40e-04       BAD
> FAIL
> 
> Beyond understanding what's going on, it would also be nice to know if there is any way to perform just that one single test for debugging purposes.

You can run the script directly, i.e. ./113-leapsecond.

BTW, if you are running the tests from a script like %check in rpm
build rather than a manual check, it's recommended to set the
CLKNETSIM_RANDOM_SEED variable to some constant value to get
deterministic results and avoid random failures. There is a compromise
between sensitivity, reliability and time required for testing, so
occassional failures are expected.

-- 
Miroslav Lichvar

-- 
To unsubscribe email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "unsubscribe" in the subject.
For help email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "help" in the subject.
Trouble?  Email listmaster@xxxxxxxxxxxxxxxxxxxx.


Mail converted by MHonArc 2.6.19+ http://listengine.tuxfamily.org/