7.2-PRERELEASE: make process waiting indefinitely
Oleg V. Nauman
oleg at opentransfer.com
Wed Apr 15 12:03:08 UTC 2009
Quoting Ronald Klop <ronald-freebsd8 at klop.yi.org>:
> On Tue, 14 Apr 2009 15:04:11 +0200, Oleg V. Nauman
> <oleg at opentransfer.com> wrote:
>
>>
>> Have experienced today strange issue with my fresh RELENG_7 (
>> sources from yesterday April 13, userland and kernel in sync ) -
>> sometimes some processes stop running without any visible reason.
>> Have seen it twice today during KDE compilation - make process just
>> waiting for something while nothing else compiles or some other
>> way prevents this process from running ( no SIGSTOP performed, no
>> Ctrl-S performed on console). Well it possible related to ( or
>> triggered by ) new ports compilation feature for SMP machines (
>> most of KDE port marked as MAKE_JOBS_SAFE=yes ) but anyway this
>> stuck process behavior is strange. kill -SIGCONT not helps, but it
>> killable ( Ctrl-C helps at least )
>> Well some info related to this process:
>> procstat -t 63472 output:
>>
>> PID TID COMM TDNAME CPU PRI STATE WCHAN
>> 63472 100059 make - 1 92 sleep wait
>>
>> procstat -kk 63472 output:
>>
>> PID TID COMM TDNAME KSTACK
>> 63472 100059 make - mi_switch+0x2c8
>> sleepq_switch+0xd9 sleepq_catch_signals+0x239 sleepq_wait_sig+0x14
>> _sleep+0x307 kern_wait+0xa36 wait4+0x3b syscall+0x2b3
>> Xint0x80_syscall+0x20
>>
>> Some related sysctls output:
>>
>> kern.smp.cpus: 2
>> kern.smp.disabled: 0
>> kern.smp.active: 1
>>
>> uname -msr output:
>> FreeBSD 7.2-PRERELEASE i386
>
> It sounds like
> http://www.mail-archive.com/freebsd-stable@freebsd.org/msg102628.html .
> There are some tips in that mailthread about debugging it.
Thank you. Looks like it was the same issue.
>
> Ronald.
More information about the freebsd-stable
mailing list