make == fork bomb ???

Andreas Maus maus at badphish.dyndns.org
Fri Apr 1 11:00:37 PST 2005


On Fri, Apr 01, 2005 at 12:14:57AM +0200, Andreas Maus wrote:
Hi Max, hi *!

So I couldn't wait till weekend and so I tried the portage tree yesterday
(around 2300 UTC) and made a buildworld && installworld...

> I will checkout and try the new version this weekend.
> At this time is just a big annoyance but not critical and
> I haven't seen sideffects now.
Nope. This doesn't really fix this problem.
make clean forks like hell:

vectra# make clean
Cannot fork: Resource temporarily unavailable
"/usr/ports/Mk/bsd.port.mk", line 1979: warning: "/usr/bin/id -u" returned non-zero status
Cannot fork: Resource temporarily unavailable
"/usr/ports/Mk/bsd.port.mk", line 2057: warning: "/usr/sbin/pkg_info -P 2>/dev/null | /usr/bin/sed -e 's/.*: //'" returned non-zero status
Cannot fork: Resource temporarily unavailable
"/usr/ports/Mk/bsd.port.mk", line 2594: warning: "/sbin/sysctl -n kern.argmax" returned non-zero status
Cannot fork: Resource temporarily unavailable
"/usr/ports/Mk/bsd.port.mk", line 1979: warning: "/usr/bin/id -u" returned non-zero status
Cannot fork: Resource temporarily unavailable
"/usr/ports/Mk/bsd.port.mk", line 2057: warning: "/usr/sbin/pkg_info -P 2>/dev/null | /usr/bin/sed -e 's/.*: //'" returned non-zero status
Cannot fork: Resource temporarily unavailable
"/usr/ports/Mk/bsd.port.mk", line 2594: warning: "/sbin/sysctl -n kern.argmax" returned non-zero status
Cannot fork: Resource temporarily unavailable
===>  Cleaning for unzip-5.52_1
===>  Cleaning for libiconv-1.9.2_1
===>  Cleaning for gettext-0.14.1
===>  Cleaning for gmake-3.80_2
===>  Cleaning for libtool-1.3.5_2
WARNING, USING DRAGONFLY OVERRIDE /usr/dfports/devel/libtool15
cd /usr/dfports/devel/libtool15 && make clean
===>  Cleaning for libtool-1.5.14

dmesg now complains that the maximum number of files are exceeded:

vectra# dmesg | tail
installed MI handler for int 7
installed MI handler for int 0
ad0: 57241MB <ST360014A> [116301/16/63] at ata0-master UDMA33
acd0: CDROM <HITACHI CDR-8435> at ata1-master PIO4
Mounting root from ufs:/dev/ad0s1a
kern.maxfiles limit exceeded by uid 0, please see tuning(7).
maxproc limit exceeded by uid 0, please see tuning(7) and login.conf(5).
maxproc limit exceeded by uid 0, please see tuning(7) and login.conf(5).
maxproc limit exceeded by uid 0, please see tuning(7) and login.conf(5).
maxproc limit exceeded by uid 0, please see tuning(7) and login.conf(5).

kern.maxfiles is:

vectra# sysctl kern.maxfiles
kern.maxfiles: 2024

Furthermore now I am unable to install a package because make is no longer
able to resolve dependencies:

vectra# pwd
/usr/ports/textproc/sgmlformat
vectra# make install 
[... snipp ...]
===>  Installing for sgmlformat-1.7_2
===>   sgmlformat-1.7_2 depends on executable: nsgmls - not found
===>    Verifying install for nsgmls in /usr/ports/textproc/jade
===>  Vulnerability check disabled, database not found
=> jade-1.2.1.tar.gz doesn't seem to exist in /usr/ports/distfiles/.
make: don't know how to make INSTALLS_DEPENDS. Stop
=> Couldn't fetch it - please try to retrieve this
=> port manually into /usr/ports/distfiles/ and try again.
*** Error code 1

Stop in /usr/ports/textproc/jade.
*** Error code 1

Stop in /usr/ports/textproc/sgmlformat.

Well I can track the dependencies down and install them manually
but this is not really a solution.

Any thoughts and ideas are welcome!

Andreas.

P.S.: I vaguely remember a similar problem years ago in FreeBSD (about 2002 or so)
where make had problems with circular dependencies. But I guess this is not the problem
in this case, isn't it?

-- 
Out of the darkness a voice spoke unto me, saying "smile, things could
be worse". So I smiled, and so, things became worse.
Attachment:
pgp00001.pgp
-------------- next part --------------
A non-text attachment was scrubbed...
Name: pgp00001.pgp
Type: application/octet-stream
Size: 481 bytes
Desc: "Description: PGP signature"
URL: <http://lists.dragonflybsd.org/pipermail/users/attachments/20050401/0844f163/attachment-0015.obj>


More information about the Users mailing list