You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
ppc64le job with ubuntu-rolling started failing when running run-on-arch-install.sh #160
Hello and thank you for this action. I started seeing failure in our downstream ppc64le using this action and wondering if the error means anything to you because I cannot make sense of it. I am not an expert in this architecture. Hope you could help or point us to someone/somewhere. 🙏
We are seeing similar issue with building our C project building on aarch64/ubuntu 20.04 and aarch64/ubuntu 22.04, which was working fine around Jan 7th.
We are facing the following error:
make[2]: *** [<builtin>: ldblib.o] Segmentation fault (core dumped)
make[2]: Leaving directory '/home/runner/work/valkey/valkey/deps/lua/src'
make[1]: *** [Makefile:102: lua] Error 2
make[1]: Leaving directory '/home/runner/work/valkey/valkey/deps'
make: [Makefile:415: persist-settings] Error 2 (ignored)
CC adlist.o
In file included from adlist.c:34:
zmalloc.h:50:10: fatal error: jemalloc/jemalloc.h: No such file or directory
50 | #include <jemalloc/jemalloc.h>
| ^~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make: *** [Makefile:467: adlist.o] Error 1
make: Leaving directory '/home/runner/work/valkey/valkey/src'
Error: The process '/home/runner/work/_actions/uraimo/run-on-arch-action/v2/src/run-on-arch.sh' failed with exit code 2
Hi all, following the discussion ongoing at tonistiigi/binfmt#215 it looks like that updating QEMU to 9.2.2 should fix the issue, I'm testing it on a branch and keeping track of the results that everyone is reporting.
As for the issue itself, I think it can be described as "sporadic segmentation faults", since even with a specific test case from CRoaring I cannot reliably replicate the issue every time.
If this works I will release a new major release, 3.0.0, since this is a pretty big update from the QEMU 7.2.0 that is in use right now through multiarch/qemu-user-static (that has not been updated for 2 years) and I don't want to introduce regressions for people using older OS releases and that could be tracking 2.x.
I'm also doing some tests with 7.0.0-7 since it's available and there are reports of 7.0.0 non being affected by this issue.
Reacted by AdamN, Julien Jerphanion, Clément Robert and Kleis Auke Wolthuizen
I can confirm that using QEMU 9.2.2 from tonistiigi/binfmt:latest fixes this. Tested this by pinning this GitHub action to 51a6e8b, see commit: kleisauke/sharp@ffd30a6.
Activity
martin-g commentedon Jan 24, 2025
Maybe pgbackrest/pgbackrest@7effca0 helps ?
TST: Use image for exotic platforms
pllim commentedon Jan 24, 2025
Thank you for your reply. I tried it in astropy/astropy#17666 but I get
Segmentation fault (core dumped). 😬martin-g commentedon Jan 28, 2025
No idea ...
You may try this workaround - pyenv/pyenv#2239 (comment)
TST: Use image for exotic platforms
roshkhatri commentedon Feb 13, 2025
We are seeing similar issue with building our C project building on
aarch64/ubuntu 20.04andaarch64/ubuntu 22.04, which was working fine around Jan 7th.We are facing the following error:
more logs can be found here: https://github.com/valkey-io/valkey/actions/runs/13295059183/job/37124939339
I have no idea why this would fail without any changes in code.
Please let me know if know any workaround that would resolve this.
Also, Thank you so much for this action, it has been very helpful.
EDIT: We just moved to github hosted arm runner which works for the project. Thanks!
anr2me commentedon Feb 27, 2025
May be an issue with QEMU, similar to tonistiigi/binfmt#215
pllim commentedon Feb 27, 2025
uraimo commentedon Feb 28, 2025
Hi all, following the discussion ongoing at tonistiigi/binfmt#215 it looks like that updating QEMU to 9.2.2 should fix the issue, I'm testing it on a branch and keeping track of the results that everyone is reporting.
As for the issue itself, I think it can be described as "sporadic segmentation faults", since even with a specific test case from CRoaring I cannot reliably replicate the issue every time.
If this works I will release a new major release, 3.0.0, since this is a pretty big update from the QEMU 7.2.0 that is in use right now through multiarch/qemu-user-static (that has not been updated for 2 years) and I don't want to introduce regressions for people using older OS releases and that could be tracking 2.x.
I'm also doing some tests with 7.0.0-7 since it's available and there are reports of 7.0.0 non being affected by this issue.
CI: Pin `uraimo/run-on-arch-action` to a specific version
CI: Pin `uraimo/run-on-arch-action` to a specific version
kleisauke commentedon Feb 28, 2025
I can confirm that using QEMU 9.2.2 from
tonistiigi/binfmt:latestfixes this. Tested this by pinning this GitHub action to 51a6e8b, see commit: kleisauke/sharp@ffd30a6.Update issue #160 test
uraimo commentedon Feb 28, 2025
3.0.0 has been published.
CI: Upgrade `uraimo/run-on-arch-action` to v3
CI: Upgrade `uraimo/run-on-arch-action` to v3 (#4340)
pllim commentedon Feb 28, 2025
Thank you! Looks promising over at astropy/astropy#17827
Update to v3: uraimo/run-on-arch-action@v3
Update to v3: uraimo/run-on-arch-action@v3
Update to v3: uraimo/run-on-arch-action@v3
Update to v3: uraimo/run-on-arch-action@v3
anr2me commentedon Mar 12, 2025
It seems the segfault issue is coming back again
Some one post about it yesterday at tonistiigi/binfmt#215 (comment)
[CI] Refresh the RISC-V runner emulation
[CI] Refresh the RISC-V runner emulation