Re: GLibc hang on in gethostbyname

[Florian Weimer <fweimer at redhat dot com>]

* David Ignjić:

> Hello all,
> Can you help me with  what can be root problem of this deadlock:
> #0  __lll_lock_wait_private () at
> ../nptl/sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:95
> #1  0x00007f1be8ee48b5 in _L_lock_34 () from /lib64/libc.so.6
> #2  0x00007f1be8ee46f3 in gethostbyname (name=name@entry=0xadf0db0 "
> xxxx.xxxxxx.com") at ../nss/getXXbyYY.c:115
> #3  0x00007f1be1316363 in XS_Socket_inet_aton (my_perl=0x1379010,
> cv=<optimized out>) at Socket.xs:595
> #4  0x00007f1bea1e441f in Perl_pp_entersub (my_perl=0x1379010) at
> pp_hot.c:2778
> #5  0x00007f1bea1dcb96 in Perl_runops_standard (my_perl=0x1379010) at
> run.c:41
> #6  0x00007f1bea179922 in S_run_body (oldscope=<optimized out>,
> my_perl=<optimized out>) at perl.c:2397
> #7  perl_run (my_perl=0x1379010) at perl.c:2320
> #8  0x0000000000400ce9 in main (argc=3, argv=0x7ffc14e23d08,
> env=0x7ffc14e23d28) at perlmain.c:120
>
> Its happens in production 1-2 times in week process hang there.
> System is centos 7 and glibc version glibc-2.17-260.el7_6.6.x86_64

Is the main program multi-threaded, and does the hang happen in a forked
subprocess?

Thanks,
Florian