MariaDB reproducibility issue

  • Done
  • quality assurance status badge
Details
4 participants
  • Ludovic Courtès
  • Christopher Baines
  • Marius Bakke
  • Josh
Owner
unassigned
Submitted by
Josh
Severity
normal
J
J
Josh wrote on 27 Jan 2020 15:23
(address . bug-guix@gnu.org)
f48564ea-f58e-41e1-a29d-faaa68ee1d8f@www.fastmail.com
Hi Guix,
I ran into this issue when building mariadb 10.1.38. I've attached the
last 300 lines of the log. Thanks
Attachment: mariadb-check.log
M
M
Marius Bakke wrote on 27 Jan 2020 20:21
87pnf4so1u.fsf@devup.no
Josh <randomenduser@fastmail.com> writes:

Toggle quote (5 lines)
> Hi Guix,
>
> I ran into this issue when building mariadb 10.1.38. I've attached the
> last 300 lines of the log. Thanks

I can reproduce this failure by checking out Guix 1.0.1 in a "time
machine" and trying to build MariaDB. The problem is that the failing test
expects the current time to be earlier than 2020-01-21 15:32:22.

This was properly fixed in 10.1.39[0], at least until the year 2038.

Unfortunately there is little we can do about it now. If you are
hitting this problem while trying to install Guix 1.0.1, I recommend to
either:

1) enable binary substitutes
2) set the system clock to some time last year while building MariaDB
3) install a minimal configuration that does not pull in MariaDB, and
reconfigure with the "full" configuration once you have 'guix pull'd.

Will that work for you?

Thanks for the report, and sorry for the inconvenience!

-----BEGIN PGP SIGNATURE-----

iQEzBAEBCgAdFiEEu7At3yzq9qgNHeZDoqBt8qM6VPoFAl4vOD0ACgkQoqBt8qM6
VPonoQgAknVOpA2VHfuWo6rwOf1WNARc7PyknReSwZ70yM/+LDRbeCeiFVha1acx
C3GGv/VDLluQV86+JvOcK52LZsPam/N53aUAeXPCnMyZzH5pQTlmJ3emHWJqUsYp
Whq5nItaVbba7OinKSSBKjTWlDxf7NIQGhISg5Va/3vAvjs2SXgv1Q2xRg6NiqH1
Bok8iQl9ULPZD7p2O14XH+pxxxtjN2lbYOIjOZtviZiNWI3nRKnyG2hHA6vyDCpZ
nS5o9E/JZ9FJuKd99yNyB10hkO4uTb7S5A1JV02IsAufjNjUwI08i7j6lq6wMsLi
l2BBJrpo9KZAUohxk7QtzhuqOml9Dw==
=ItLZ
-----END PGP SIGNATURE-----

L
L
Ludovic Courtès wrote on 24 Jan 2021 17:12
(name . Marius Bakke)(address . mbakke@fastmail.com)
87k0s271qx.fsf@gnu.org
Hi!

Marius Bakke <mbakke@fastmail.com> skribis:

Toggle quote (11 lines)
> Josh <randomenduser@fastmail.com> writes:
>
>> Hi Guix,
>>
>> I ran into this issue when building mariadb 10.1.38. I've attached the
>> last 300 lines of the log. Thanks
>
> I can reproduce this failure by checking out Guix 1.0.1 in a "time
> machine" and trying to build MariaDB. The problem is that the failing test
> expects the current time to be earlier than 2020-01-21 15:32:22.

I was trying to run:

guix time-machine --commit=0791437 -- build mariadb

where commit 0791437f972caa7e48de91ad5cb150a614f617c2 is from Jan. 2019,
and stumbled upon the test failure that Josh reported.

Following your advice on IRC, Marius, I built that derivation,
/gnu/store/p2d7i5258vi0rd9ydbpr9c1vb3sxcz6h-mariadb-10.1.37.drv, on a
machine whose clocked I had switched back to 2018… and it worked!

(I used one of the berlin build machines, so now there are substitutes
for this particular derivation.)

Thanks!

Ludo’.
C
C
Christopher Baines wrote on 1 Nov 2023 10:53
(address . 39310-done@debbugs.gnu.org)
87bkcd94sz.fsf@cbaines.net
Ludovic Courtès <ludo@gnu.org> writes:

Toggle quote (29 lines)
> Hi!
>
> Marius Bakke <mbakke@fastmail.com> skribis:
>
>> Josh <randomenduser@fastmail.com> writes:
>>
>>> Hi Guix,
>>>
>>> I ran into this issue when building mariadb 10.1.38. I've attached the
>>> last 300 lines of the log. Thanks
>>
>> I can reproduce this failure by checking out Guix 1.0.1 in a "time
>> machine" and trying to build MariaDB. The problem is that the failing test
>> expects the current time to be earlier than 2020-01-21 15:32:22.
>
> I was trying to run:
>
> guix time-machine --commit=0791437 -- build mariadb
>
> where commit 0791437f972caa7e48de91ad5cb150a614f617c2 is from Jan. 2019,
> and stumbled upon the test failure that Josh reported.
>
> Following your advice on IRC, Marius, I built that derivation,
> /gnu/store/p2d7i5258vi0rd9ydbpr9c1vb3sxcz6h-mariadb-10.1.37.drv, on a
> machine whose clocked I had switched back to 2018… and it worked!
>
> (I used one of the berlin build machines, so now there are substitutes
> for this particular derivation.)

This seems like this issue is addressed at least, and guix challenge
says mariadb is reproducible, so I'm going to mark this as done:

→ guix challenge --diff=simple mariadb

1 store items were analyzed:
- 1 (100.0%) were identical
- 0 (0.0%) differed
- 0 (0.0%) were inconclusive
-----BEGIN PGP SIGNATURE-----
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=TXmS
-----END PGP SIGNATURE-----

Closed
?