GNU bug report logs - #57868
GStreamer gst_gstsystemclock test fails on CI for i686

Previous Next

To reply to this bug, email your comments to 57868 AT debbugs.gnu.org.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Marius Bakke <marius <at> gnu.org>
To: bug-guix <at> gnu.org
Subject: GStreamer gst_gstsystemclock test fails on CI for i686
Date: Sat, 17 Sep 2022 00:30:13 +0200

Hi,

On i686-linux, the gstreamer gst_gstsystemclock test is failing:

--8<---------------cut here---------------start------------->8---
 41/108 gst_gstsystemclock            FAIL            1.26s   exit status 2
>>> GST_PLUGIN_SYSTEM_PATH_1_0='' CK_DEFAULT_TIMEOUT=600 GST_STATE_IGNORE_ELEMENTS='' MALLOC_PERTURB_=230 GST_PLUGIN_SCANNER_1_0=/tmp/guix-build-gstreamer-1.20.3.drv-0/build/libs/gst/helpers/gst-plugin-scanner GST_PLUGIN_LOADING_WHITELIST=gstreamer GST_REGISTRY=/tmp/guix-build-gstreamer-1.20.3.drv-0/build/tests/check/gst_gstsystemclock.registry GST_PLUGIN_PATH_1_0=/tmp/guix-build-gstreamer-1.20.3.drv-0/build /tmp/guix-build-gstreamer-1.20.3.drv-0/build/tests/check/gst_gstsystemclock
????????????????????????????????????? ?  ?????????????????????????????????????
stdout:
Running suite(s): GstSystemClock
75%: Checks: 8, Failures: 0, Errors: 2
../gstreamer-1.20.3/tests/check/gst/gstsystemclock.c:263:E:waiting:test_stress_cleanup_unschedule:0: (after this point) Received signal 5 (Trace/breakpoint trap)
../gstreamer-1.20.3/tests/check/gst/gstsystemclock.c:263:E:waiting:test_stress_reschedule:0: (after this point) Received signal 5 (Trace/breakpoint trap)
Check suite gst_systemclock ran in 1.236s (tests failed: 2)
stderr:

(gst_gstsystemclock:5546): GLib-ERROR **: 22:57:16.217: creating thread 'wait': Error creating thread: Resource temporarily unavailable

(gst_gstsystemclock:6027): GLib-ERROR **: 22:57:16.340: creating thread 'wait': Error creating thread: Resource temporarily unavailable
??????????????????????????????????????????????????????????????????????????????
--8<---------------cut here---------------end--------------->8---

Full log output here:

  https://ci.guix.gnu.org/build/1305526/log/raw

I'm not able to reproduce this locally.  Any idea what might be going on
here?  Parallelism issue?

Note: we have disabled these tests on i686 for a long time, but they
were recently enabled again on 'staging'.  I'll re-apply that hunk to
disable the test, but wanted to have an issue to link to.

Message #8 received at 57868 <at> debbugs.gnu.org (full text, mbox):

From: Nicolas Graves <ngraves <at> ngraves.fr>
To: Marius Bakke <marius <at> gnu.org>, 57868 <at> debbugs.gnu.org
Subject: Re: bug#57868: GStreamer gst_gstsystemclock test fails on CI for i686
Date: Mon, 25 Dec 2023 19:00:19 +0100

Hi Marius,

I think a similar issue also happens on x86_64. For some reason, the
binary is available through guix weather, but I'm unable to build the
package locally (with --check). Can you confirm it's not just on my
side?
 
I'm going to send a patch which basically adds this phase for all
platforms after this email, I'll be using this locally. 

Cheers,
Nicolas 


On 2022-09-17 00:30, Marius Bakke wrote:

> Hi,
>
> On i686-linux, the gstreamer gst_gstsystemclock test is failing:
>
> --8<---------------cut here---------------start------------->8---
>  41/108 gst_gstsystemclock            FAIL            1.26s   exit status 2
>>>> GST_PLUGIN_SYSTEM_PATH_1_0='' CK_DEFAULT_TIMEOUT=600 GST_STATE_IGNORE_ELEMENTS='' MALLOC_PERTURB_=230 GST_PLUGIN_SCANNER_1_0=/tmp/guix-build-gstreamer-1.20.3.drv-0/build/libs/gst/helpers/gst-plugin-scanner GST_PLUGIN_LOADING_WHITELIST=gstreamer GST_REGISTRY=/tmp/guix-build-gstreamer-1.20.3.drv-0/build/tests/check/gst_gstsystemclock.registry GST_PLUGIN_PATH_1_0=/tmp/guix-build-gstreamer-1.20.3.drv-0/build /tmp/guix-build-gstreamer-1.20.3.drv-0/build/tests/check/gst_gstsystemclock
> ????????????????????????????????????? ?  ?????????????????????????????????????
> stdout:
> Running suite(s): GstSystemClock
> 75%: Checks: 8, Failures: 0, Errors: 2
> ../gstreamer-1.20.3/tests/check/gst/gstsystemclock.c:263:E:waiting:test_stress_cleanup_unschedule:0: (after this point) Received signal 5 (Trace/breakpoint trap)
> ../gstreamer-1.20.3/tests/check/gst/gstsystemclock.c:263:E:waiting:test_stress_reschedule:0: (after this point) Received signal 5 (Trace/breakpoint trap)
> Check suite gst_systemclock ran in 1.236s (tests failed: 2)
> stderr:
>
> (gst_gstsystemclock:5546): GLib-ERROR **: 22:57:16.217: creating thread 'wait': Error creating thread: Resource temporarily unavailable
>
> (gst_gstsystemclock:6027): GLib-ERROR **: 22:57:16.340: creating thread 'wait': Error creating thread: Resource temporarily unavailable
> ??????????????????????????????????????????????????????????????????????????????
> --8<---------------cut here---------------end--------------->8---
>
> Full log output here:
>
>   https://ci.guix.gnu.org/build/1305526/log/raw
>
> I'm not able to reproduce this locally.  Any idea what might be going on
> here?  Parallelism issue?
>
> Note: we have disabled these tests on i686 for a long time, but they
> were recently enabled again on 'staging'.  I'll re-apply that hunk to
> disable the test, but wanted to have an issue to link to.
>
>
>

-- 
Best regards,
Nicolas Graves

Message #11 received at 57868 <at> debbugs.gnu.org (full text, mbox):

From: Nicolas Graves <ngraves <at> ngraves.fr>
To: 57868 <at> debbugs.gnu.org
Cc: ngraves <at> ngraves.fr, marius <at> gnu.org
Subject: [PATCH] gnu: gstreamer: Phase disable-systemclock-test on all
 platforms.
Date: Mon, 25 Dec 2023 19:02:51 +0100

* gnu/packages/gstreamer.scm (gstreamer):
[arguments]: Enable phase disable-systemclock-test on all platforms.

Change-Id: I346934acab66f7a33a9cea6500c8557509b05d95
---
 gnu/packages/gstreamer.scm | 19 +++++++------------
 1 file changed, 7 insertions(+), 12 deletions(-)

diff --git a/gnu/packages/gstreamer.scm b/gnu/packages/gstreamer.scm
index 71f8e3453a..3608576830 100644
--- a/gnu/packages/gstreamer.scm
+++ b/gnu/packages/gstreamer.scm
@@ -477,20 +477,15 @@ (define-public gstreamer
            #:phases
            #~(modify-phases %standard-phases
                #$@%common-gstreamer-phases
-               #$@(if (string-prefix? "i686" (or (%current-target-system)
-                                                 (%current-system)))
-                      ;; FIXME: These tests consistently fail in the Guix CI:
-                      ;;   https://issues.guix.gnu.org/57868
-                      '((add-after 'unpack 'disable-systemclock-test
-                          (lambda _
-                            (substitute* "tests/check/gst/gstsystemclock.c"
-                              (("tcase_add_test \\(tc_chain, \
+               (add-after 'unpack 'disable-systemclock-test
+                 (lambda _
+                   (substitute* "tests/check/gst/gstsystemclock.c"
+                     (("tcase_add_test \\(tc_chain, \
 test_stress_cleanup_unschedule.*")
-                               "")
-                              (("tcase_add_test \\(tc_chain, \
+                      "")
+                     (("tcase_add_test \\(tc_chain, \
 test_stress_reschedule.*")
-                               "")))))
-                      '())
+                      ""))))
                (add-after 'patch-shebangs 'do-not-capture-python
                  (lambda _
                    ;; The patch-source-shebangs phase causes the following build
-- 
2.41.0

Message #14 received at 57868 <at> debbugs.gnu.org (full text, mbox):

From: Nicolas Graves <ngraves <at> ngraves.fr>
To: Marius Bakke <marius <at> gnu.org>, 57868 <at> debbugs.gnu.org
Subject: Re: bug#57868: GStreamer gst_gstsystemclock test fails on CI for i686
Date: Tue, 26 Dec 2023 10:08:13 +0100

On 2023-12-25 19:00, Nicolas Graves wrote:

> Hi Marius,
>
> I think a similar issue also happens on x86_64. For some reason, the
> binary is available through guix weather, but I'm unable to build the
> package locally (with --check). Can you confirm it's not just on my
> side?

Sorry it's not the gst_gstsystemclock but rather the
libs_gstnetclientclock that fails consistently. You can ignore the patch
sent. 

>  
> I'm going to send a patch which basically adds this phase for all
> platforms after this email, I'll be using this locally. 
>
> Cheers,
> Nicolas 
>
>
> On 2022-09-17 00:30, Marius Bakke wrote:
>
>> Hi,
>>
>> On i686-linux, the gstreamer gst_gstsystemclock test is failing:
>>
>> --8<---------------cut here---------------start------------->8---
>>  41/108 gst_gstsystemclock            FAIL            1.26s   exit status 2
>>>>> GST_PLUGIN_SYSTEM_PATH_1_0='' CK_DEFAULT_TIMEOUT=600 GST_STATE_IGNORE_ELEMENTS='' MALLOC_PERTURB_=230 GST_PLUGIN_SCANNER_1_0=/tmp/guix-build-gstreamer-1.20.3.drv-0/build/libs/gst/helpers/gst-plugin-scanner GST_PLUGIN_LOADING_WHITELIST=gstreamer GST_REGISTRY=/tmp/guix-build-gstreamer-1.20.3.drv-0/build/tests/check/gst_gstsystemclock.registry GST_PLUGIN_PATH_1_0=/tmp/guix-build-gstreamer-1.20.3.drv-0/build /tmp/guix-build-gstreamer-1.20.3.drv-0/build/tests/check/gst_gstsystemclock
>> ????????????????????????????????????? ?  ?????????????????????????????????????
>> stdout:
>> Running suite(s): GstSystemClock
>> 75%: Checks: 8, Failures: 0, Errors: 2
>> ../gstreamer-1.20.3/tests/check/gst/gstsystemclock.c:263:E:waiting:test_stress_cleanup_unschedule:0: (after this point) Received signal 5 (Trace/breakpoint trap)
>> ../gstreamer-1.20.3/tests/check/gst/gstsystemclock.c:263:E:waiting:test_stress_reschedule:0: (after this point) Received signal 5 (Trace/breakpoint trap)
>> Check suite gst_systemclock ran in 1.236s (tests failed: 2)
>> stderr:
>>
>> (gst_gstsystemclock:5546): GLib-ERROR **: 22:57:16.217: creating thread 'wait': Error creating thread: Resource temporarily unavailable
>>
>> (gst_gstsystemclock:6027): GLib-ERROR **: 22:57:16.340: creating thread 'wait': Error creating thread: Resource temporarily unavailable
>> ??????????????????????????????????????????????????????????????????????????????
>> --8<---------------cut here---------------end--------------->8---
>>
>> Full log output here:
>>
>>   https://ci.guix.gnu.org/build/1305526/log/raw
>>
>> I'm not able to reproduce this locally.  Any idea what might be going on
>> here?  Parallelism issue?
>>
>> Note: we have disabled these tests on i686 for a long time, but they
>> were recently enabled again on 'staging'.  I'll re-apply that hunk to
>> disable the test, but wanted to have an issue to link to.
>>
>>
>>

-- 
Best regards,
Nicolas Graves

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.