Bug#928921: reprotest: kernel too old SIGSEGV

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Bug#928921: reprotest: kernel too old SIGSEGV

Dmitry Bogatov-3

Package: reprotest
Version: 0.7.8
Severity: normal

Hello. I am working on setting up CI for my packages, and one of them
fail "reprotest" phase with error: "FATAL: kernel too old"[1][2].

I grepped upstream sources for "kernel too old" message and found
nothing. Regular build is successful. Any suggestions, what may be wrong
and how to fix it?

On another package I get "bad system call" error[2].

 [^1] https://salsa.debian.org/debian/bglibs/-/jobs/170367
 [^2] https://salsa.debian.org/debian/bcron/-/jobs/170195
 [^3] https://salsa.debian.org/debian/gdbm/-/jobs/170585
--
        Note, that I send and fetch email in batch, once every 24 hours.
                 If matter is urgent, try https://t.me/kaction
                                                                             --

attachment0 (849 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Bug#928921: reprotest: kernel too old SIGSEGV

Vagrant Cascadian-9
On 2019-05-13, Dmitry Bogatov wrote:
> Hello. I am working on setting up CI for my packages, and one of them
> fail "reprotest" phase with error: "FATAL: kernel too old"[1][2].

This is probably due to the kernel variation using "setarch", which will
cause uname to report a 2.6 kernel, which is too old for some things...

You should be able to disable it with --variations=+all,-kernel


live well,
  vagrant

signature.asc (233 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Bug#928921: reprotest: kernel too old SIGSEGV

Dmitry Bogatov-3

control: retitle -1 reprotest: unable to chown

[2019-05-12 22:12] Vagrant Cascadian <[hidden email]>
> On 2019-05-13, Dmitry Bogatov wrote:
> > Hello. I am working on setting up CI for my packages, and one of them
> > fail "reprotest" phase with error: "FATAL: kernel too old"[1][2].
>
> This is probably due to the kernel variation using "setarch", which will
> cause uname to report a 2.6 kernel, which is too old for some things...
>
> You should be able to disable it with --variations=+all,-kernel

Thank you very much. It did solve the issue. But while setting up CI for
another package, I encountered another issue:

        install: fatal: unable to chown /tmp/reprotest.1SX4Bu/build-experiment-1/debian/qmail/var/lib/qmail: invalid argument

Package builds fine, both on CI and locally, and content


 [full log](https://salsa.debian.org/kaction/qmail/-/jobs/188411)



--
        Note, that I send and fetch email in batch, once every 24 hours.
                 If matter is urgent, try https://t.me/kaction
                                                                             --

Reply | Threaded
Open this post in threaded view
|

Bug#928921: reprotest: kernel too old SIGSEGV

Jonas Meurer
Control: subscribe -1

Hey,

Dmitry Bogatov <[hidden email]> wrote:
> But while setting up CI for another package, I encountered another issue:
>
> install: fatal: unable to chown /tmp/reprotest.1SX4Bu/build-experiment-1/debian/qmail/var/lib/qmail: invalid argument
>
> Package builds fine, both on CI and locally, and content

I run into the exactly same issue (chown erroring iwth 'Invalid
argument') in the reprotest job of salsa-ci on my package mailman-suite:

https://salsa.debian.org/mailman-team/mailman-suite/-/jobs/192550

chown: modification du propriƩtaire de
'debian/mailman3-web/var/lib/mailman3': Argument invalide

The package has 'chown list:list' for several files in debian/rules.

Trying to reproduce and debug it locally (with the help of IƱaki Malerba
and Agustin Henze - kudos to them), we found out that 'unshare -r' is
the problem here. Removing the '-r' argument to unshare makes the chown
work (and the package build) again.

I wonder whether it's a good idea to map the current user+group to root
in reprotest at all, as it definitely breaks chown at build time.

I'm looking forward to your (the maintainers') take on this ;)

Cheers
 jonas


signature.asc (849 bytes) Download Attachment