Bug#890430: make: warning: -jN forced in submake: disabling jobserver mode.

[Andreas Beckmann]

On Fri, 16 Feb 2018 14:34:40 +0100 Aurelien Jarno 
wrote:
> This problem also affects the build daemons, the build of the linux
> package is much longer now on the architectures where make-dfsg does not
> FTBFS. The linux package indeed builds in parallel by using the
> MAKEFLAGS variable. I guess other packages are also affected. I am
> therefore raising the severity of this bug, we do not want such
> regression in testing.

This also affects openjdk-10 (and probably the other openjdk-* as well):
https://bugs.debian.org/891573
There the result was a unlimited-parallel build (something like
make -j), I noticed because the machine became unresponsive for
interactive use because the load suddenly went above 400 and it started
to swap ...


Andreas

Bug#890430: make: warning: -jN forced in submake: disabling jobserver mode.

[Aurelien Jarno]

control: severity -1 serious

On 2018-02-14 16:13, Manoj Srivastava wrote:
> Hi,
> 
> While you do not explicitly mention it, I suspect you have -jN
>  specified in  MAKEFLAGS variable. his is a known issue in make 4.2.1;
>  the fix is in upstream commit
>  0c5a9f9b92af1634dc60fa21e9ac86ed50e5d595.
> 
> Please also see https://savannah.gnu.org/bugs/?48274
> 

This problem also affects the build daemons, the build of the linux
package is much longer now on the architectures where make-dfsg does not
FTBFS. The linux package indeed builds in parallel by using the
MAKEFLAGS variable. I guess other packages are also affected. I am
therefore raising the severity of this bug, we do not want such
regression in testing.

Aurelien

-- 
Aurelien Jarno  GPG: 4096R/1DDD8C9B
aurel...@aurel32.net http://www.aurel32.net


signature.asc
Description: PGP signature

Bug#890430: make: warning: -jN forced in submake: disabling jobserver mode.

[Manoj Srivastava]

Hi,

Unfortunately, the fix is not as simple as just  cherry-picking
 the related commit (there are earlier changes, and at least one part of
 the commit is later reverted).

Since we already have a RC bug on make, I’ll wait until there is
 a fix for that (since current HEAD reorganizes the source tree, which
 makes carrying patches ickier)

manoj
-- 
"Cogito ergo I'm right and you're wrong." Blair Houghton
Manoj Srivastava  
4096R/C5779A1C E37E 5EC5 2A01 DA25 AD20  05B6 CF48 9438 C577 9A1C


smime.p7s
Description: S/MIME cryptographic signature

Bug#890430: make: warning: -jN forced in submake: disabling jobserver mode.

[Manoj Srivastava]

Hi,

While you do not explicitly mention it, I suspect you have -jN
 specified in  MAKEFLAGS variable. his is a known issue in make 4.2.1;
 the fix is in upstream commit
 0c5a9f9b92af1634dc60fa21e9ac86ed50e5d595.

Please also see https://savannah.gnu.org/bugs/?48274

manoj
-- 
When asked, "If you find so much that is unworthy of reverence in the
United States, then why do you live here?" Mencken replied, "Why do men
go to zoos?"
Manoj Srivastava  
4096R/C5779A1C E37E 5EC5 2A01 DA25 AD20  05B6 CF48 9438 C577 9A1C


smime.p7s
Description: S/MIME cryptographic signature

Bug#890430: make: warning: -jN forced in submake: disabling jobserver mode.

[Sven Joachim]

Package: make
Version: 4.2.1-1

The following change in make 4.2, mentioned in the NEWS file:

,
| * The amount of parallelism can be determined by querying MAKEFLAGS, even when
|   the job server is enabled (previously MAKEFLAGS would always contain only
|   "-j", with no number, when job server was enabled).
`

triggers the following warning here when building the ncurses package,
shown several dozen times:

,
| make[2]: warning: -jN forced in submake: disabling jobserver mode.
`

More importantly, the build time has increased considerably, this is
what I get on a laptop with four rather lame cores:

,
| % LANG=C; time dpkg-buildpackage -b > /dev/null 2>&1
| dpkg-buildpackage -b > /dev/null 2>&1  962.98s user 134.50s system 238% cpu 
7:39.25 total
`

In contrast, building with make 4.1.9 gives a much better CPU utilization:

,
| % LANG=C; time dpkg-buildpackage -b > /dev/null 2>&1 
| dpkg-buildpackage -b > /dev/null 2>&1  983.30s user 136.62s system 338% cpu 
5:31.28 total
`


-- System Information:
Debian Release: buster/sid
  APT prefers unstable
  APT policy: (500, 'unstable'), (101, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.14.18-pavil (SMP w/4 CPU cores)
Locale: LANG=de_DE.UTF-8, LC_CTYPE=de_DE.UTF-8 (charmap=UTF-8), 
LANGUAGE=de_DE.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Versions of packages make depends on:
ii  libc6  2.26-6

make recommends no packages.

Versions of packages make suggests:
pn  make-doc  

-- no debconf information