1. Forum
  2. Usenet
  3. LINUX.DEBIAN.BUGS.RC

  • Bug#1103360: firefox-esr: FTBFS in testing/arm64: make[6]: *** [/build/

    From Mike Hommey@21:1/5 to Lucas Nussbaum on Wed Apr 16 23:10:01 2025
    On Wed, Apr 16, 2025 at 08:40:14PM +0200, Lucas Nussbaum wrote:
    Source: firefox-esr
    Version: 128.9.0esr-2
    Severity: serious
    Justification: FTBFS
    Tags: trixie sid ftbfs
    User: [email protected]
    Usertags: ftbfs-20250416 ftbfs-trixie

    Hi,

    During a rebuild of all packages in testing (trixie), your package failed
    to build on arm64.

    The most relevant part of the log says: (signal: 9, SIGKILL: kill)

    This was OOM killed. IOW, not enough memory on the builder you used.

    Is this something that runs on buildds or some other machines?

    Mike

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Santiago Vila@21:1/5 to All on Wed Apr 16 23:40:01 2025
    El 16/4/25 a las 23:07, Mike Hommey escribió:
    The most relevant part of the log says: (signal: 9, SIGKILL: kill)

    This was OOM killed. IOW, not enough memory on the builder you used.

    Is this something that runs on buildds or some other machines?

    Hi. This is an archive rebuild for QA purposes, not the build which
    happens in the official buildds.

    I see by looking at the build log that the machine has 16GB of RAM
    and probably no swap. This is more or less the memory required to build firefox-esr on a machine with 1 CPU, so I'm not surprised that the
    build is killed because not having enough memory.

    [ I know how much memory firefox-esr needs because I collect
    memory usage statistics during my own archive rebuilds ]

    More CPUs usually makes the memory consumption higher.

    Does the machine really have 0 cores, as shown in the build log?

    DC-System-Info: 0 cores (), 16007536 kB RAM

    Can anybody see the parallelism being used in the build log?

    Maybe this should not be RC, but in either case it might be worth
    to be investigated more.

    Thanks.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Lucas Nussbaum@21:1/5 to Santiago Vila on Thu Apr 17 00:00:01 2025
    On 16/04/25 at 23:30 +0200, Santiago Vila wrote:
    El 16/4/25 a las 23:07, Mike Hommey escribi�:
    The most relevant part of the log says: (signal: 9, SIGKILL: kill)

    This was OOM killed. IOW, not enough memory on the builder you used.

    Is this something that runs on buildds or some other machines?

    Hi. This is an archive rebuild for QA purposes, not the build which
    happens in the official buildds.

    I see by looking at the build log that the machine has 16GB of RAM
    and probably no swap. This is more or less the memory required to build firefox-esr on a machine with 1 CPU, so I'm not surprised that the
    build is killed because not having enough memory.

    [ I know how much memory firefox-esr needs because I collect
    memory usage statistics during my own archive rebuilds ]

    More CPUs usually makes the memory consumption higher.

    Does the machine really have 0 cores, as shown in the build log?

    No it has four cores. It's a bug in the script that generates the
    header.

    DC-System-Info: 0 cores (), 16007536 kB RAM

    Can anybody see the parallelism being used in the build log?

    Maybe this should not be RC,

    Yes

    but in either case it might be worth
    to be investigated more.

    I'll do my next ARM rebuild with more RAM.

    Lucas

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)