Discussion:
State of riscv64
(too old to reply)
Cyril Brulebois
2024-12-26 22:00:01 UTC
Permalink
Hi,

I'd like to get a sense of whether the following situation on riscv64 is
just pure bad luck or if that's something that we should expect to run
into in a regular fashion:

1: gcc-14 (11d 11h 30m, rv-manda-02), graph-tool (4d 19h 33m, rv-osuosl-03), gcc-snapshot (4d 15h 1m, rv-osuosl-04), kicad (11h 8m, rv-manda-01), mpich (7h 14m, rv-manda-04), pyside6 (5h 26m, rv-manda-03), freecad (5h 9m, rv-osuosl-05), jpeg-xl (4h 19m, rv-osuosl-02)

A number of those packages take several-to-many days to build. The other
ones “only” take many hours instead.

I wouldn't usually worry about this too much, but it seems riscv64
became a full release architecture in the meanwhile (even if I don't
recall seeing any announcement), and missing builds block migration


https://salsa.debian.org/release-team/britney2/-/commit/593e01d703966ba8602eeff7594b0553900ea84f


Cheers,
--
Cyril Brulebois (***@debian.org) <https://debamax.com/>
D-I release manager -- Release team member -- Freelance Consultant
Sebastian Ramacher
2024-12-31 12:20:01 UTC
Permalink
Hi
Post by Cyril Brulebois
I'd like to get a sense of whether the following situation on riscv64 is
just pure bad luck or if that's something that we should expect to run
1: gcc-14 (11d 11h 30m, rv-manda-02), graph-tool (4d 19h 33m, rv-osuosl-03), gcc-snapshot (4d 15h 1m, rv-osuosl-04), kicad (11h 8m, rv-manda-01), mpich (7h 14m, rv-manda-04), pyside6 (5h 26m, rv-manda-03), freecad (5h 9m, rv-osuosl-05), jpeg-xl (4h 19m, rv-osuosl-02)
To some extent this is bad luck with uploads of packages that take a
long time to build. From my experience with the last couple of
transitions, riscv64 was reasonable fast to work though the rebuilds.

Also Loading Image... suggests that the
backlog is usually not too bad. Note that the current peak of packages
in Needs-Build is due to low priority rebuilds of hundreds of packages
with outdated Built-Using.

Cheers
--
Sebastian Ramacher
Cyril Brulebois
2025-01-03 17:10:01 UTC
Permalink
Post by Sebastian Ramacher
Post by Cyril Brulebois
I'd like to get a sense of whether the following situation on riscv64 is
just pure bad luck or if that's something that we should expect to run
1: gcc-14 (11d 11h 30m, rv-manda-02), graph-tool (4d 19h 33m, rv-osuosl-03), gcc-snapshot (4d 15h 1m, rv-osuosl-04), kicad (11h 8m, rv-manda-01), mpich (7h 14m, rv-manda-04), pyside6 (5h 26m, rv-manda-03), freecad (5h 9m, rv-osuosl-05), jpeg-xl (4h 19m, rv-osuosl-02)
To some extent this is bad luck with uploads of packages that take a
long time to build. From my experience with the last couple of
transitions, riscv64 was reasonable fast to work though the rebuilds.
Also https://buildd.debian.org/stats/riscv64.png suggests that the
backlog is usually not too bad. Note that the current peak of packages
in Needs-Build is due to low priority rebuilds of hundreds of packages
with outdated Built-Using.
Copy all, thank you.


Cheers,
--
Cyril Brulebois (***@debian.org) <https://debamax.com/>
D-I release manager -- Release team member -- Freelance Consultant
Loading...