Tests time out when building in slow environments, like emulated
s390x in Fedora copr. Bump up the test timeout

Signed-off-by: Cole Robinson <crobinso@redhat.com>
---
 libvirt.spec.in | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/libvirt.spec.in b/libvirt.spec.in
index 0a8b0ebad4..1731aa8bd9 100644
--- a/libvirt.spec.in
+++ b/libvirt.spec.in
@@ -1288,7 +1288,9 @@ mv $RPM_BUILD_ROOT%{_datadir}/systemtap/tapset/libvirt_qemu_probes.stp \
 %endif
 
 %check
-VIR_TEST_DEBUG=1 %meson_test --no-suite syntax-check
+# Building on slow archs, like emulated s390x in Fedora copr, requires
+# raising the test timeout
+VIR_TEST_DEBUG=1 %meson_test --no-suite syntax-check --timeout-multiplier 10
 
 %post libs
 %if 0%{?rhel} == 7
-- 
2.29.2

On Thu, Jan 21, 2021 at 4:59 PM Cole Robinson <crobinso@redhat.com> wrote:
>
> Tests time out when building in slow environments, like emulated
> s390x in Fedora copr. Bump up the test timeout
>
> Signed-off-by: Cole Robinson <crobinso@redhat.com>
> ---
>  libvirt.spec.in | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/libvirt.spec.in b/libvirt.spec.in
> index 0a8b0ebad4..1731aa8bd9 100644
> --- a/libvirt.spec.in
> +++ b/libvirt.spec.in
> @@ -1288,7 +1288,9 @@ mv $RPM_BUILD_ROOT%{_datadir}/systemtap/tapset/libvirt_qemu_probes.stp \
>  %endif
>
>  %check
> -VIR_TEST_DEBUG=1 %meson_test --no-suite syntax-check
> +# Building on slow archs, like emulated s390x in Fedora copr, requires
> +# raising the test timeout
> +VIR_TEST_DEBUG=1 %meson_test --no-suite syntax-check --timeout-multiplier 10
>
>  %post libs
>  %if 0%{?rhel} == 7
> --
> 2.29.2
>

I have to do this for slow hardware too, so I'm fine with it.

Reviewed-by: Neal Gompa <ngompa13@gmail.com>


-- 
真実はいつも一つ！/ Always, there's only one truth!

On 1/21/21 10:55 PM, Cole Robinson wrote:
> Tests time out when building in slow environments, like emulated
> s390x in Fedora copr. Bump up the test timeout
> 
> Signed-off-by: Cole Robinson <crobinso@redhat.com>
> ---
>   libvirt.spec.in | 4 +++-
>   1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/libvirt.spec.in b/libvirt.spec.in
> index 0a8b0ebad4..1731aa8bd9 100644
> --- a/libvirt.spec.in
> +++ b/libvirt.spec.in
> @@ -1288,7 +1288,9 @@ mv $RPM_BUILD_ROOT%{_datadir}/systemtap/tapset/libvirt_qemu_probes.stp \
>   %endif
>   
>   %check
> -VIR_TEST_DEBUG=1 %meson_test --no-suite syntax-check
> +# Building on slow archs, like emulated s390x in Fedora copr, requires
> +# raising the test timeout
> +VIR_TEST_DEBUG=1 %meson_test --no-suite syntax-check --timeout-multiplier 10
>   
>   %post libs
>   %if 0%{?rhel} == 7
> 

Looks like we have consensus here. In the past, we've increased the 
timeout for qemuxml2argvtest:

https://gitlab.com/libvirt/libvirt/-/commit/94146c9d2ba

Would it make sense finally admit that not everybody is running top 
shelf CPU and increase the timeout even for 'ninja test'? I mean, set 
timeout to a bigger value in tests/meson.build and drop special casing 
of one test?

Michal

On 1/25/21 5:54 AM, Michal Privoznik wrote:
> On 1/21/21 10:55 PM, Cole Robinson wrote:
>> Tests time out when building in slow environments, like emulated
>> s390x in Fedora copr. Bump up the test timeout
>>
>> Signed-off-by: Cole Robinson <crobinso@redhat.com>
>> ---
>>   libvirt.spec.in | 4 +++-
>>   1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/libvirt.spec.in b/libvirt.spec.in
>> index 0a8b0ebad4..1731aa8bd9 100644
>> --- a/libvirt.spec.in
>> +++ b/libvirt.spec.in
>> @@ -1288,7 +1288,9 @@ mv
>> $RPM_BUILD_ROOT%{_datadir}/systemtap/tapset/libvirt_qemu_probes.stp \
>>   %endif
>>     %check
>> -VIR_TEST_DEBUG=1 %meson_test --no-suite syntax-check
>> +# Building on slow archs, like emulated s390x in Fedora copr, requires
>> +# raising the test timeout
>> +VIR_TEST_DEBUG=1 %meson_test --no-suite syntax-check
>> --timeout-multiplier 10
>>     %post libs
>>   %if 0%{?rhel} == 7
>>
> 
> Looks like we have consensus here. In the past, we've increased the
> timeout for qemuxml2argvtest:
> 
> https://gitlab.com/libvirt/libvirt/-/commit/94146c9d2ba
> 
> Would it make sense finally admit that not everybody is running top
> shelf CPU and increase the timeout even for 'ninja test'? I mean, set
> timeout to a bigger value in tests/meson.build and drop special casing
> of one test?
> 

+1 from me. In my case the slow test was virschematest

- Cole

On Thu, 2021-01-21 at 16:55 -0500, Cole Robinson wrote:
>  %check
> -VIR_TEST_DEBUG=1 %meson_test --no-suite syntax-check
> +# Building on slow archs, like emulated s390x in Fedora copr, requires
> +# raising the test timeout
> +VIR_TEST_DEBUG=1 %meson_test --no-suite syntax-check --timeout-multiplier 10

Debian has been doing the same thing for a while, even though in that
case it's to cope with slow native hardware rather than emulation:

  https://salsa.debian.org/libvirt-team/libvirt/-/blob/debian/6.9.0-3/debian/rules#L170-179

So

  Reviewed-by: Andrea Bolognani <abologna@redhat.com>

from my side, but please leave the message on the list for a few days
to give other people the opportunity to voice any concerns they might
have.

-- 
Andrea Bolognani / Red Hat / Virtualization

On 1/22/21 2:46 AM, Andrea Bolognani wrote:
> On Thu, 2021-01-21 at 16:55 -0500, Cole Robinson wrote:
>>   %check
>> -VIR_TEST_DEBUG=1 %meson_test --no-suite syntax-check
>> +# Building on slow archs, like emulated s390x in Fedora copr, requires
>> +# raising the test timeout
>> +VIR_TEST_DEBUG=1 %meson_test --no-suite syntax-check --timeout-multiplier 10
> 
> Debian has been doing the same thing for a while, even though in that
> case it's to cope with slow native hardware rather than emulation:
> 
>    https://salsa.debian.org/libvirt-team/libvirt/-/blob/debian/6.9.0-3/debian/rules#L170-179

FYI, I also use the timeout multiplier in the SUSE libvirt package to avoid test 
failures in OBS, particularly for non-x86 architectures. A value of 5 has worked 
fine thus far.

Regards,
Jim