Bug#872635: util-linux: FTBFS on armel: test failure

[Andy Simpkins]

I have had a look at this as part of the Cambridge BSP 2019-03-09

I am able to reproduce this 'bug', on multiple architectures  the
following is copy/paste from buster on my AMD64 laptop :-p

Simply running the test by hand
Assuming you have a working / reliable resolver / untainted cache then
the command succeeds:


../util-linux-2.33.1/tests/ts/utmp$ last -f wtmp-ipv6.LE -d  root
IPv6 root a.root-servers.n Wed Aug 28 21:30 - 21:40  (00:10)

wtmp-ipv6.LE begins Wed Aug 28 21:30:40 2013




../util-linux-2.33.1/tests/ts/utmp$ last -f wtmp-ipv6.LE -w -d root
IPv6 root a.root-servers.net Wed Aug 28 21:30 - 21:40  (00:10)

wtmp-ipv6.LE begins Wed Aug 28 21:30:40 2013




../util-linux-2.33.1/tests/ts/utmp$ last -f wtmp-ipv6.LE -a -d root
IPv6 root Wed Aug 28 21:30 - 21:40  (00:10)
a.root-servers.net

wtmp-ipv6.LE begins Wed Aug 28 21:30:40 2013







causing a failure can be done simply by unplugging the machine from the
network, thus...

../util-linux-2.33.1/tests/ts/utmp$ last -f wtmp-ipv6.LE -d root
IPv6 root dns-server   Wed Aug 28 21:30 - 21:40  (00:10)

wtmp-ipv6.LE begins Wed Aug 28 21:30:40 2013



../util-linux-2.33.1/tests/ts/utmp$ last -f wtmp-ipv6.LE -w -d root
IPv6 root dns-server   Wed Aug 28 21:30 - 21:40  (00:10)

wtmp-ipv6.LE begins Wed Aug 28 21:30:40 2013



/util-linux-2.33.1/tests/ts/utmp$ last -f wtmp-ipv6.LE -a -d root
IPv6 root Wed Aug 28 21:30 - 21:40  (00:10) dns-server

wtmp-ipv6.LE begins Wed Aug 28 21:30:40 2013



From this I conclude that the test itself is poor - it is assuming that
there is a consistently good network / resolver for the duration of the
test, something that can not be assumed to always be true. Theoretically
glitches can happen anywhere, at any time.

Question.  What is the purpose of these 3 specific tests?
 If we are confirming that 'last' is formatting output in the manor
specified by the switches then the tests are successful:  the reported
output may contain EITHER a.root-servers.net XOR dns-server
If we are testing that the lookup has happened then it is perfectly
acceptable that it will not be possible due to a transitory network
failure.  this does not mean that last itself is not working correctly
only that this test did not PASS (i.e. !PASS is not the same as FAIL)

Maybe remove/disable the test, or add retries? Maybe add detection for a
timeout failure and simply return a different value to warn about this?



signature.asc
Description: OpenPGP digital signature

Bug#872635: util-linux: FTBFS on armel: test failure

[Michael Biebl]

Forwarding this bug report to the debian-arm porters list so not only
Aurelien and Steve are aware of the issue.

Would be great if you can take a look at
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=872635
and help us find a solution.

On Sat, 19 Aug 2017 17:10:13 +0200 Cyril Brulebois  wrote:
> Source: util-linux
> Version: 2.29.2-4
> Severity: serious
> Justification: FTBFS
> 
> Hi,
> 
> Your package FTBFS on armel, hopefully I've managed to pick the right part:
> | ~~~ dns long ~~~
> | ~~~ dns host last ~~~
> |  FAILED (utmp/last-dns)
> | = script: /<>/tests/ts/utmp/last =
> | = OUTPUT =
> |  1  ~~~ dns short ~~~
> |  2  IPv4 root dns-server   Wed Aug 28 13:00 - 14:00 
>  (01:00)
> |  3  
> |  4  wtmp-a begins Wed Aug 28 03:00:00 2013
> |  5  ~~~ dns long ~~~
> |  6  IPv4 root dns-server   Wed Aug 28 13:00 - 14:00 
>  (01:00)
> |  7  
> |  8  wtmp-a begins Wed Aug 28 03:00:00 2013
> |  9  ~~~ dns host last ~~~
> | 10  IPv4 root Wed Aug 28 13:00 - 14:00  (01:00) 
> dns-server
> | 11  
> | 12  wtmp-a begins Wed Aug 28 03:00:00 2013
> | = EXPECTED ===
> |  1  ~~~ dns short ~~~
> |  2  IPv4 root a.root-servers.n Wed Aug 28 13:00 - 14:00 
>  (01:00)
> |  3  
> |  4  wtmp-a begins Wed Aug 28 03:00:00 2013
> |  5  ~~~ dns long ~~~
> |  6  IPv4 root a.root-servers.net Wed Aug 28 13:00 - 
> 14:00  (01:00)
> |  7  
> |  8  wtmp-a begins Wed Aug 28 03:00:00 2013
> |  9  ~~~ dns host last ~~~
> | 10  IPv4 root Wed Aug 28 13:00 - 14:00  (01:00) 
> a.root-servers.net
> | 11  
> | 12  wtmp-a begins Wed Aug 28 03:00:00 2013
> | = O/E diff ===
> | --- /<>/tests/output/utmp/last-dns 2017-08-17 
> 13:37:15.211650587 +
> | +++ /<>/tests/expected/utmp/last-dns   2016-11-02 
> 12:03:35.569215943 +
> | @@ -1,12 +1,12 @@
> |  ~~~ dns short ~~~
> | -IPv4 root dns-server   Wed Aug 28 13:00 - 14:00  (01:00)
> | +IPv4 root a.root-servers.n Wed Aug 28 13:00 - 14:00  (01:00)
> |  
> |  wtmp-a begins Wed Aug 28 03:00:00 2013
> |  ~~~ dns long ~~~
> | -IPv4 root dns-server   Wed Aug 28 13:00 - 14:00  (01:00)
> | +IPv4 root a.root-servers.net Wed Aug 28 13:00 - 14:00  (01:00)
> |  
> |  wtmp-a begins Wed Aug 28 03:00:00 2013
> |  ~~~ dns host last ~~~
> | -IPv4 root Wed Aug 28 13:00 - 14:00  (01:00) dns-server
> | +IPv4 root Wed Aug 28 13:00 - 14:00  (01:00) 
> a.root-servers.net
> |  
> |  wtmp-a begins Wed Aug 28 03:00:00 2013
> | ==
> | : nodns   ...~~~ basic output ~~~
> | ~~~ include system ~~~

-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?



signature.asc
Description: OpenPGP digital signature

Bug#872635: util-linux: FTBFS on armel: test failure

[Andreas Henriksson]

Hello,

On Sat, Aug 19, 2017 at 05:10:13PM +0200, Cyril Brulebois wrote:
> Source: util-linux
> Version: 2.29.2-4
> Severity: serious
> Justification: FTBFS
> 
> Hi,
> 
> Your package FTBFS on armel, hopefully I've managed to pick the right part:
> | ~~~ dns long ~~~
> | ~~~ dns host last ~~~
> |  FAILED (utmp/last-dns)
[...]

This is a well known issue with the ARM Ltd hosted builders and not armel
specific.

I haven't been able to find a way to reproduce this outside the ARM Ltd
hosted buildds.

The solution is simply to reschedule the build and hope it ends up
on a buildd not hosted at ARM Ltd.

The problem is basically that first this[1] condition passes, then
subsequent[2][3][4] (identical) calls fails (at ARM Ltd):

[1]: http://sources.debian.net/src/util-linux/2.29.2-4/tests/ts/utmp/last/#L41
[2]: http://sources.debian.net/src/util-linux/2.29.2-4/tests/ts/utmp/last/#L45
[3]: http://sources.debian.net/src/util-linux/2.29.2-4/tests/ts/utmp/last/#L48
[4]: http://sources.debian.net/src/util-linux/2.29.2-4/tests/ts/utmp/last/#L51

This has previously been discussed on IRC and the people responsible
for buildd hosting at ARM Ltd has been unable to identify what's
causing it.

AFAIK there's a possibility to blacklist buildds somehow. Please reassign
this bug report to where/who-ever can help out with blacklisting ARM Ltd
buildds for util-linux builds.


Regards,
Andreas Henriksson

Bug#872635: util-linux: FTBFS on armel: test failure

[Cyril Brulebois]

Source: util-linux
Version: 2.29.2-4
Severity: serious
Justification: FTBFS

Hi,

Your package FTBFS on armel, hopefully I've managed to pick the right part:
| ~~~ dns long ~~~
| ~~~ dns host last ~~~
|  FAILED (utmp/last-dns)
| = script: /<>/tests/ts/utmp/last =
| = OUTPUT =
|  1~~~ dns short ~~~
|  2IPv4 root dns-server   Wed Aug 28 13:00 - 14:00 
 (01:00)
|  3
|  4wtmp-a begins Wed Aug 28 03:00:00 2013
|  5~~~ dns long ~~~
|  6IPv4 root dns-server   Wed Aug 28 13:00 - 14:00 
 (01:00)
|  7
|  8wtmp-a begins Wed Aug 28 03:00:00 2013
|  9~~~ dns host last ~~~
| 10IPv4 root Wed Aug 28 13:00 - 14:00  (01:00) 
dns-server
| 11
| 12wtmp-a begins Wed Aug 28 03:00:00 2013
| = EXPECTED ===
|  1~~~ dns short ~~~
|  2IPv4 root a.root-servers.n Wed Aug 28 13:00 - 14:00 
 (01:00)
|  3
|  4wtmp-a begins Wed Aug 28 03:00:00 2013
|  5~~~ dns long ~~~
|  6IPv4 root a.root-servers.net Wed Aug 28 13:00 - 
14:00  (01:00)
|  7
|  8wtmp-a begins Wed Aug 28 03:00:00 2013
|  9~~~ dns host last ~~~
| 10IPv4 root Wed Aug 28 13:00 - 14:00  (01:00) 
a.root-servers.net
| 11
| 12wtmp-a begins Wed Aug 28 03:00:00 2013
| = O/E diff ===
| --- /<>/tests/output/utmp/last-dns   2017-08-17 
13:37:15.211650587 +
| +++ /<>/tests/expected/utmp/last-dns 2016-11-02 
12:03:35.569215943 +
| @@ -1,12 +1,12 @@
|  ~~~ dns short ~~~
| -IPv4 root dns-server   Wed Aug 28 13:00 - 14:00  (01:00)
| +IPv4 root a.root-servers.n Wed Aug 28 13:00 - 14:00  (01:00)
|  
|  wtmp-a begins Wed Aug 28 03:00:00 2013
|  ~~~ dns long ~~~
| -IPv4 root dns-server   Wed Aug 28 13:00 - 14:00  (01:00)
| +IPv4 root a.root-servers.net Wed Aug 28 13:00 - 14:00  (01:00)
|  
|  wtmp-a begins Wed Aug 28 03:00:00 2013
|  ~~~ dns host last ~~~
| -IPv4 root Wed Aug 28 13:00 - 14:00  (01:00) dns-server
| +IPv4 root Wed Aug 28 13:00 - 14:00  (01:00) 
a.root-servers.net
|  
|  wtmp-a begins Wed Aug 28 03:00:00 2013
| ==
| : nodns   ...~~~ basic output ~~~
| ~~~ include system ~~~
| ~~~ show ip ~~~
| ~~~ no host ~~~
| ~~~ full times shortopt ~~~
| ~~~ full times longopt ~~~
| ~~~ no time ~~~
| ~~~ iso-8601 time ~~~
| ~~~ since and until ~~~
| ~~~ present ~~~
|  OK
|... FAILED (1 from 2 sub-tests)
| = script: /<>/tests/ts/utmp/last =
| = OUTPUT =
| = EXPECTED ===
| = O/E diff ===
| ==
| 
|  script: /<>/tests/ts/utmp/last-ipv6
| sub dir: /<>/tests/ts/utmp
| top dir: /<>/tests
|self: /<>/tests/ts/utmp
|   test name: last-ipv6
|   test desc: last ipv6
|   component: utmp
|   namespace: utmp/last-ipv6
| verbose: yes
|  output: /<>/tests/output/utmp/last-ipv6
|valgrind: /<>/tests/output/utmp/last-ipv6.vgdump
|expected: /<>/tests/expected/utmp/last-ipv6
|  mountpoint: /<>/tests/output/utmp/last-ipv6-mnt
| 
|  utmp: last ipv6  ...
| : dns ...~~~ dns short ~~~
| ~~~ dns long ~~~
[…]
| -
|   1 tests of 186 FAILED
| -
| Makefile:12092: recipe for target 'check-local-tests' failed
| make[5]: *** [check-local-tests] Error 1
| make[5]: Leaving directory '/<>'
| Makefile:11734: recipe for target 'check-am' failed
| make[4]: *** [check-am] Error 2
| make[4]: Leaving directory '/<>'
| Makefile:11445: recipe for target 'check-recursive' failed
| make[3]: *** [check-recursive] Error 1
| make[3]: Leaving directory '/<>'
| Makefile:11738: recipe for target 'check' failed
| make[2]: *** [check] Error 2
| make[2]: Leaving directory '/<>'
| dh_auto_test: make -j1 check VERBOSE=1 returned exit code 2
| debian/rules:159: recipe for target 'override_dh_auto_test' failed
| make[1]: *** [override_dh_auto_test] Error 2
| make[1]: Leaving directory '/<>'
| debian/rules:47: recipe for target 'build-arch' failed
| make: *** [build-arch] Error 2
| dpkg-buildpackage: error: debian/rules build-arch gave error exit status 2

Full build log:
  
https://buildd.debian.org/status/fetch.php?pkg=util-linux=armel=2.29.2-4=1502977052=0


KiBi.