pdns_server --list-modules dumps core on OpenBSD #2422

paultheweird · 2015-04-10T16:12:24Z

When running pdns_server --list-modules I get a core dump on OpenBSD. This is PowerDNS 3.4.3 (port not updated in OpenBSD ports tree yet):

[weerd@pom] $ pdns_server --list-modules         
Apr 10 18:10:49 Reading random entropy from '/dev/urandom'
Apr 10 18:10:49 Loading '/usr/local/lib/powerdns/pdns/libgeobackend.so'
Apr 10 18:10:49 Loading '/usr/local/lib/powerdns/pdns/libgsqlite3backend.so'
Modules available:
geo
gsqlite3
pdns_server(3822) in free(): error: chunk is already free 0xb065ba27d00
Abort trap (core dumped)

Running gdb on the core file I get the following backtrace:

(gdb) bt
#0  0x00000b061c1b991a in kill () at <stdin>:2
#1  0x00000b061c1f30d9 in abort () at /usr/src/lib/libc/stdlib/abort.c:53
#2  0x00000b061c1ef7f0 in wrterror (msg=0xb061c2fc476 "chunk is already free", p=0xb065ba27d00)
    at /usr/src/lib/libc/stdlib/malloc.c:281
#3  0x00000b061c1ef8e6 in find_chunknum (d=Variable "d" is not available.
) at /usr/src/lib/libc/stdlib/malloc.c:1005
#4  0x00000b061c1f0aff in ofree (p=0xb065ba27d00) at /usr/src/lib/libc/stdlib/malloc.c:1240
#5  0x00000b061c1f0cee in free (ptr=0xb065ba27d00) at /usr/src/lib/libc/stdlib/malloc.c:1282
#6  0x00000b0374e75bba in GeoBackend::queueNSRecords () from /usr/local/sbin/pdns_server
#7  0x00000b061c1d5d49 in __cxa_finalize (dso=0x0) at /usr/src/lib/libc/stdlib/atexit.c:141
#8  0x00000b061c1c7f1e in exit (status=99) at /usr/src/lib/libc/stdlib/exit.c:57
#9  0x00000b0374f42384 in main () from /usr/local/sbin/pdns_server
Current language:  auto; currently asm

The text was updated successfully, but these errors were encountered:

fobser · 2015-04-11T10:33:32Z

With the geo backend disabled and the bind backend enabled I see a similar crash:

$ pdns_server --list-modules
Apr 11 12:01:41 Reading random entropy from '/dev/urandom'
Apr 11 12:01:41 Loading '/usr/local/lib/powerdns/pdns/libbindbackend.so'
Apr 11 12:01:41 Loading '/usr/local/lib/powerdns/pdns/libgsqlite3backend.so'
Modules available:
bind
gsqlite3
Bus error (core dumped)

(gdb) bt
#0  0x00001b17a3d7bb36 in Bind2Backend::safeRemoveBBDomainInfo ()
   from /usr/local/sbin/pdns_server
#1  0x00001b1a18c8ef59 in __cxa_finalize (dso=0x0)
    at /usr/src/lib/libc/stdlib/atexit.c:142
#2  0x00001b1a18c6ac4e in exit (status=99)
    at /usr/src/lib/libc/stdlib/exit.c:57
#3  0x00001b17a3e57ef4 in main () from /usr/local/sbin/pdns_server
(gdb)

With only the sqlite backend enabled --list-modules works fine.
A limited test of the bind backend shows that it's working fine, this seems to be only an issue with --list-modules.

cmouse · 2015-04-11T11:13:09Z

Any chance you could run valgrind on this?

fobser · 2015-04-11T11:25:23Z

On 04/11/15 13:13, Aki Tuomi wrote:

Any chance you could run valgrind on this?

Not at this time, openbsd doesn't have valgrind, yet.

cmouse · 2015-04-11T11:28:41Z

http://ports.su/devel/valgrind ?

fobser · 2015-04-12T11:32:39Z

Oh, heh, I didn't know he valgrind to work yet.
This is with the geo backend enabled again:

$ valgrind /usr/local/sbin/pdns_server --list-modules 
==7972== Memcheck, a memory error detector
==7972== Copyright (C) 2002-2013, and GNU GPL'd, by Julian Seward et al.
==7972== Using Valgrind-3.10.1 and LibVEX; rerun with -h for copyright info
==7972== Command: /usr/local/sbin/pdns_server --list-modules
==7972== 
==7972== Invalid read of size 8
==7972==    at 0x72D83B6: pthread_self (rthread.c:279)
==7972==    by 0x72D8464: _rthread_dl_lock (rthread.c:669)
==7972==    by 0x72D8050: _rthread_init (rthread.c:214)
==7972==    by 0x72D83C6: pthread_self (rthread.c:276)
==7972==    by 0x72DB13A: _rthread_mutex_lock (rthread_sync.c:88)
==7972==    by 0x72D7374: pthread_once (rthread_once.c:24)
==7972==    by 0x5272102: __gthread_once (gthr-default.h:575)
==7972==    by 0x5272102: std::locale::_S_initialize() (locale_init.cc:265)
==7972==    by 0x527240C: std::locale::locale() (locale_init.cc:214)
==7972==    by 0x52576C0: basic_streambuf (streambuf:462)
==7972==    by 0x52576C0: stdio_sync_filebuf (stdio_sync_filebuf.h:72)
==7972==    by 0x52576C0: std::ios_base::Init::Init() (ios_init.cc:89)
==7972==    by 0x583B3BA: __static_initialization_and_destruction_0(int, int) (iostream:77)
==7972==    by 0x583B416: global constructors keyed to libs_serialization_src_xml_oarchive.cpp_1FD5A050_A88192F8 (xml_oarchive.cpp:33)
==7972==    by 0x57F7BFD: ??? (in /usr/local/lib/libboost_serialization-mt.so.5.1)
==7972==  Address 0x10 is not stack'd, malloc'd or (recently) free'd
==7972== 
==7972== 
==7972== Process terminating with default action of signal 11 (SIGSEGV): dumping core
==7972==  Access not within mapped region at address 0x10
==7972==    at 0x72D83B6: pthread_self (rthread.c:279)
==7972==    by 0x72D8464: _rthread_dl_lock (rthread.c:669)
==7972==    by 0x72D8050: _rthread_init (rthread.c:214)
==7972==    by 0x72D83C6: pthread_self (rthread.c:276)
==7972==    by 0x72DB13A: _rthread_mutex_lock (rthread_sync.c:88)
==7972==    by 0x72D7374: pthread_once (rthread_once.c:24)
==7972==    by 0x5272102: __gthread_once (gthr-default.h:575)
==7972==    by 0x5272102: std::locale::_S_initialize() (locale_init.cc:265)
==7972==    by 0x527240C: std::locale::locale() (locale_init.cc:214)
==7972==    by 0x52576C0: basic_streambuf (streambuf:462)
==7972==    by 0x52576C0: stdio_sync_filebuf (stdio_sync_filebuf.h:72)
==7972==    by 0x52576C0: std::ios_base::Init::Init() (ios_init.cc:89)
==7972==    by 0x583B3BA: __static_initialization_and_destruction_0(int, int) (iostream:77)
==7972==    by 0x583B416: global constructors keyed to libs_serialization_src_xml_oarchive.cpp_1FD5A050_A88192F8 (xml_oarchive.cpp:33)
==7972==    by 0x57F7BFD: ??? (in /usr/local/lib/libboost_serialization-mt.so.5.1)
==7972==  If you believe this happened as a result of a stack
==7972==  overflow in your program's main thread (unlikely but
==7972==  possible), you can try to increase the size of the
==7972==  main thread stack using the --main-stacksize= flag.
==7972==  The main thread stack size used in this run was 4194304.

valgrind: m_coredump/coredump-elf.c:816 (make_elf_coredump): Assertion 'VG_(lseek)(core_fd, phdrs[idx].p_offset, VKI_SEEK_SET) == phdrs[idx].p_offset' failed.

host stacktrace:

==7972==    at 0x3803B343: ???
==7972==    by 0x8030BEFDF: ???
==7972==    by 0x38054882: ???

sched status:
  running_tid=1

Thread 1: status = VgTs_Runnable
==7972==    at 0x72D83B6: pthread_self (rthread.c:279)
==7972==    by 0x72D8464: _rthread_dl_lock (rthread.c:669)
==7972==    by 0x72D8050: _rthread_init (rthread.c:214)
==7972==    by 0x72D83C6: pthread_self (rthread.c:276)
==7972==    by 0x72DB13A: _rthread_mutex_lock (rthread_sync.c:88)
==7972==    by 0x72D7374: pthread_once (rthread_once.c:24)
==7972==    by 0x5272102: __gthread_once (gthr-default.h:575)
==7972==    by 0x5272102: std::locale::_S_initialize() (locale_init.cc:265)
==7972==    by 0x527240C: std::locale::locale() (locale_init.cc:214)
==7972==    by 0x52576C0: basic_streambuf (streambuf:462)
==7972==    by 0x52576C0: stdio_sync_filebuf (stdio_sync_filebuf.h:72)
==7972==    by 0x52576C0: std::ios_base::Init::Init() (ios_init.cc:89)
==7972==    by 0x583B3BA: __static_initialization_and_destruction_0(int, int) (iostream:77)
==7972==    by 0x583B416: global constructors keyed to libs_serialization_src_xml_oarchive.cpp_1FD5A050_A88192F8 (xml_oarchive.cpp:33)
==7972==    by 0x57F7BFD: ??? (in /usr/local/lib/libboost_serialization-mt.so.5.1)
==7972==    by 0x57E8D18: ??? (in /usr/local/lib/libboost_serialization-mt.so.5.1)

cmouse · 2015-04-12T12:25:34Z

Can you compile the code again without removing debugging symbols and see what gdb thinks. To be sure, you can add -ggdb to the CFLAGS and CXXFLAGS.

paultheweird · 2015-04-14T07:49:44Z

After updating the port to put the modules in /usr/local/lib/pdns/ (as opposed to /usr/local/lib/powerdns/pdns/), the crash seems to be gone:

[weerd@pom] $ pdns_server --list-modules
Apr 14 08:57:15 Reading random entropy from '/dev/urandom'
Apr 14 08:57:15 Loading '/usr/local/lib/pdns/libgsqlite3backend.so'
Apr 14 08:57:15 Loading '/usr/local/lib/pdns/libbindbackend.so'
Apr 14 08:57:15 Loading '/usr/local/lib/pdns/libgeobackend.so'
Apr 14 08:57:15 Loading '/usr/local/lib/pdns/libldapbackend.so'
Apr 14 08:57:15 Loading '/usr/local/lib/pdns/libgmysqlbackend.so'
Apr 14 08:57:15 Loading '/usr/local/lib/pdns/libgpgsqlbackend.so'
Modules available:
bind
geo
gmysql
gpgsql
gpgsql2
gsqlite3
ldap
[weerd@pom] $

cmouse · 2015-04-14T07:52:14Z

Did you update PKGLIBDIR when moving the modules to the other directory?

cmouse · 2015-04-14T07:52:33Z

Or module-dir in powerdns config

paultheweird · 2015-04-14T08:02:19Z

The alternative location was set with --libdir="${PREFIX}/lib/powerdns" as a configure flag (where PREFIX defaults to /usr/local/ in the OpenBSD ports tree).

cmouse · 2015-04-14T08:03:35Z

Can you test if the crash occurs if you define "module-dir= /usr/local/lib/powerdns/pdns" in pdns.conf?

paultheweird · 2015-04-14T08:06:37Z

It doesn't:

[weerd@pom] $ sudo vi pdns.conf                     
[weerd@pom] $ pdns_server --list-modules            
Apr 14 10:05:28 Reading random entropy from '/dev/urandom'
Apr 14 10:05:28 Unable to open module directory '/usr/local/lib/powerdns/pdns'
Modules available:
[weerd@pom] $ sudo mkdir -p /usr/local/lib/powerdns/pdns
[weerd@pom] $ sudo cp /usr/local/lib/pdns/* /usr/local/lib/powerdns/pdns/
[weerd@pom] $ pdns_server --list-modules                                 
Apr 14 10:05:55 Reading random entropy from '/dev/urandom'
Apr 14 10:05:55 Loading '/usr/local/lib/powerdns/pdns/libbindbackend.so'
Apr 14 10:05:55 Loading '/usr/local/lib/powerdns/pdns/libgeobackend.so'
Apr 14 10:05:55 Loading '/usr/local/lib/powerdns/pdns/libgmysqlbackend.so'
Apr 14 10:05:55 Loading '/usr/local/lib/powerdns/pdns/libgpgsqlbackend.so'
Apr 14 10:05:55 Loading '/usr/local/lib/powerdns/pdns/libgsqlite3backend.so'
Apr 14 10:05:55 Loading '/usr/local/lib/powerdns/pdns/libldapbackend.so'
Modules available:
bind
geo
gmysql
gpgsql
gpgsql2
gsqlite3
ldap
[weerd@pom] $

So I guess that rules out the module-dir part?

cmouse · 2015-04-14T08:07:36Z

did you recompile it to original problem state first? can you run grep PKGLIBDIR in some Makefile?

paultheweird · 2015-04-14T08:08:07Z

Oh, sorry, no I didn't. Let me do that now. Will report in a few minutes.

paultheweird · 2015-04-14T08:20:51Z

Confirmed that a binary with the /usr/local/lib/powerdns/pdns path configured and that same directory configured in the pdns.conf module-dir option still crashes:

[weerd@pom] $ grep ^module-dir /etc/pdns/pdns.conf 
module-dir=/usr/local/lib/powerdns/pdns
[weerd@pom] $ pdns_server --list-modules          
Apr 14 10:18:18 Reading random entropy from '/dev/urandom'
Apr 14 10:18:18 Loading '/usr/local/lib/powerdns/pdns/libgsqlite3backend.so'
Apr 14 10:18:18 Loading '/usr/local/lib/powerdns/pdns/libbindbackend.so'
Apr 14 10:18:18 Loading '/usr/local/lib/powerdns/pdns/libgeobackend.so'
Modules available:
bind
geo
gsqlite3
pdns_server(23216) in free(): error: chunk is already free 0x6f99a2c9640
Abort trap (core dumped) 
[weerd@pom] $

Also, here's the grep for PKGLIBDIR:

[weerd@pom] $ grep -r PKGLIBDIR /usr/ports/pobj/powerdns-3.4.3/pdns-3.4.3/pdns/Makefile
/usr/ports/pobj/powerdns-3.4.3/pdns-3.4.3/pdns/Makefile:        -DPKGLIBDIR=\"$(pkglibdir)\" -DLOCALSTATEDIR=\"$(socketdir)\" \
[weerd@pom] $

cmouse · 2015-04-14T08:26:53Z

um, can you grep again for pkglibdir?

paultheweird · 2015-04-14T08:28:27Z

pkglibdir = $(libdir)/pdns

and

libdir = /usr/local/lib/powerdns

cmouse · 2015-04-14T08:34:03Z

interesting that it crashes when you change that. i can't find any references in the code for PKGLIBDIR other than as default value for module-dir. did you compile with -ggdb, can you use gdb to get us a stack trace?

paultheweird · 2015-04-14T08:36:41Z

I'd have to look into that as I'm not 100% sure about the details. Will get back to you later today or tomorrow. Thanks for your interest so far!

cmouse · 2015-04-14T08:37:55Z

make clean
./configure <stuff here> CXXFLAGS="-O2 -Wall -ggdb" CFLAGS="-O2 -Wall -ggdb"
make
gdb program
r <parameter>
<after crash>
bt

paultheweird · 2015-04-14T10:39:25Z

(gdb) bt
#0  0x00000163f7b8d9ba in kill () at <stdin>:2
#1  0x00000163f7bc7839 in abort () at /usr/src/lib/libc/stdlib/abort.c:53
#2  0x00000163f7b9c1b0 in wrterror (msg=0x163f7cceeb6 "chunk is already free", p=0x163a8dc7800)
    at /usr/src/lib/libc/stdlib/malloc.c:281
#3  0x00000163f7b9c2a6 in find_chunknum (d=Variable "d" is not available.
) at /usr/src/lib/libc/stdlib/malloc.c:1010
#4  0x00000163f7b9d4cf in ofree (p=0x163a8dc7800) at /usr/src/lib/libc/stdlib/malloc.c:1245
#5  0x00000163f7b9d6be in free (ptr=0x163a8dc7800) at /usr/src/lib/libc/stdlib/malloc.c:1287
#6  0x000001613037b53a in __tcf_6 () at basic_string.h:238
#7  0x00000163f7bc3f59 in __cxa_finalize (dso=0x0) at /usr/src/lib/libc/stdlib/atexit.c:142
#8  0x00000163f7b9fc4e in exit (status=99) at /usr/src/lib/libc/stdlib/exit.c:57
#9  0x0000016130467604 in main (argc=2, argv=0x7f7fffff0a08) at receiver.cc:519
Current language:  auto; currently asm
(gdb)

cmouse · 2015-04-14T11:09:08Z

Can you run valgrind once more? Not that I expect it to reveal anything new.

fobser · 2015-04-14T12:13:33Z

I don't quite get why valgrind outputs ??? for boost, it's build with debug symbols
and not stripped.

[florian@openbsd-build:~]$ valgrind pdns_server --list-modules
==27696== Memcheck, a memory error detector
==27696== Copyright (C) 2002-2013, and GNU GPL'd, by Julian Seward et al.
==27696== Using Valgrind-3.10.1 and LibVEX; rerun with -h for copyright info
==27696== Command: pdns_server --list-modules
==27696==
==27696== Invalid read of size 8
==27696== at 0x5D813B6: pthread_self (rthread.c:279)
==27696== by 0x5D81464: _rthread_dl_lock (rthread.c:669)
==27696== by 0x5D81050: _rthread_init (rthread.c:214)
==27696== by 0x5D813C6: pthread_self (rthread.c:276)
==27696== by 0x5D8413A: _rthread_mutex_lock (rthread_sync.c:88)
==27696== by 0x5D80374: pthread_once (rthread_once.c:24)
==27696== by 0x4C9F102: __gthread_once (gthr-default.h:575)
==27696== by 0x4C9F102: std::locale::_S_initialize() (locale_init.cc:265)
==27696== by 0x4C9F40C: std::locale::locale() (locale_init.cc:214)
==27696== by 0x4C846C0: basic_streambuf (streambuf:462)
==27696== by 0x4C846C0: stdio_sync_filebuf (stdio_sync_filebuf.h:72)
==27696== by 0x4C846C0: std::ios_base::Init::Init() (ios_init.cc:89)
==27696== by 0x71E13BA: __static_initialization_and_destruction_0(int, int) (iostream:77)
==27696== by 0x71E1416: global constructors keyed to libs_serialization_src_xml_oarchive.cpp_1FD5A050_A88192F8 (xml_oarchive.cpp:33)
==27696== by 0x719DBFD: ??? (in /usr/local/lib/libboost_serialization-mt.so.5.1)
==27696== Address 0x10 is not stack'd, malloc'd or (recently) free'd
==27696==
==27696==
==27696== Process terminating with default action of signal 11 (SIGSEGV): dumping core
==27696== Access not within mapped region at address 0x10
==27696== at 0x5D813B6: pthread_self (rthread.c:279)
==27696== by 0x5D81464: _rthread_dl_lock (rthread.c:669)
==27696== by 0x5D81050: _rthread_init (rthread.c:214)
==27696== by 0x5D813C6: pthread_self (rthread.c:276)
==27696== by 0x5D8413A: _rthread_mutex_lock (rthread_sync.c:88)
==27696== by 0x5D80374: pthread_once (rthread_once.c:24)
==27696== by 0x4C9F102: __gthread_once (gthr-default.h:575)
==27696== by 0x4C9F102: std::locale::_S_initialize() (locale_init.cc:265)
==27696== by 0x4C9F40C: std::locale::locale() (locale_init.cc:214)
==27696== by 0x4C846C0: basic_streambuf (streambuf:462)
==27696== by 0x4C846C0: stdio_sync_filebuf (stdio_sync_filebuf.h:72)
==27696== by 0x4C846C0: std::ios_base::Init::Init() (ios_init.cc:89)
==27696== by 0x71E13BA: __static_initialization_and_destruction_0(int, int) (iostream:77)
==27696== by 0x71E1416: global constructors keyed to libs_serialization_src_xml_oarchive.cpp_1FD5A050_A88192F8 (xml_oarchive.cpp:33)
==27696== by 0x719DBFD: ??? (in /usr/local/lib/libboost_serialization-mt.so.5.1)
==27696== If you believe this happened as a result of a stack
==27696== overflow in your program's main thread (unlikely but
==27696== possible), you can try to increase the size of the
==27696== main thread stack using the --main-stacksize= flag.
==27696== The main thread stack size used in this run was 4194304.

valgrind: m_coredump/coredump-elf.c:816 (make_elf_coredump): Assertion 'VG_(lseek)(core_fd, phdrs[idx].p_offset, VKI_SEEK_SET) == phdrs[idx].p_offset' failed.

host stacktrace:
==27696== at 0x3803B343: ???
==27696== by 0x8030BEFDF: ???
==27696== by 0x38054882: ???

sched status:
running_tid=1

Thread 1: status = VgTs_Runnable
==27696== at 0x5D813B6: pthread_self (rthread.c:279)
==27696== by 0x5D81464: _rthread_dl_lock (rthread.c:669)
==27696== by 0x5D81050: _rthread_init (rthread.c:214)
==27696== by 0x5D813C6: pthread_self (rthread.c:276)
==27696== by 0x5D8413A: _rthread_mutex_lock (rthread_sync.c:88)
==27696== by 0x5D80374: pthread_once (rthread_once.c:24)
==27696== by 0x4C9F102: __gthread_once (gthr-default.h:575)
==27696== by 0x4C9F102: std::locale::_S_initialize() (locale_init.cc:265)
==27696== by 0x4C9F40C: std::locale::locale() (locale_init.cc:214)
==27696== by 0x4C846C0: basic_streambuf (streambuf:462)
==27696== by 0x4C846C0: stdio_sync_filebuf (stdio_sync_filebuf.h:72)
==27696== by 0x4C846C0: std::ios_base::Init::Init() (ios_init.cc:89)
==27696== by 0x71E13BA: __static_initialization_and_destruction_0(int, int) (iostream:77)
==27696== by 0x71E1416: global constructors keyed to libs_serialization_src_xml_oarchive.cpp_1FD5A050_A88192F8 (xml_oarchive.cpp:33)
==27696== by 0x719DBFD: ??? (in /usr/local/lib/libboost_serialization-mt.so.5.1)
==27696== by 0x718ED18: ??? (in /usr/local/lib/libboost_serialization-mt.so.5.1)

cmouse · 2015-04-14T16:48:48Z

Wonder if this can be due to our hardening features... can you provide the CXXFLAGS, CFLAGS and LDFLAGS from Makefile?

paultheweird · 2015-04-15T13:26:30Z

Sorry for the slow reply.

CXXFLAGS = -pthread -O2 -Wall -ggdb
CFLAGS = -O2 -Wall -ggdb
LDFLAGS = -pthread -L/usr/local/lib

I would suspect the OpenBSD hardening features before PowerDNS hardening features (since this seems an OpenBSD-only problem). Happy to be proven wrong though :)

cmouse · 2015-04-15T13:38:43Z

Those are not used here, it seems, unless the compiler itself adds some stuff by default.

paultheweird · 2015-04-15T13:40:52Z

On OpenBSD, binaries are built with several hardening features by default (pie and stack protection come to mind).

rubenk · 2015-05-05T14:54:44Z

This is fixed by 44ac597

ahupowerdns · 2015-05-05T19:13:36Z

did you test it? I didn’t.

rubenk · 2015-05-05T19:14:12Z

I did.

ahupowerdns · 2015-05-05T19:20:32Z

Thank you!

rubenk · 2015-05-05T19:21:38Z

Is this worth backporting to 3.x?

Habbie · 2015-06-16T19:41:29Z

geobackend is gone, closing this

Habbie added auth defect labels Apr 29, 2015

Habbie closed this as completed Jun 16, 2015

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

pdns_server --list-modules dumps core on OpenBSD #2422

pdns_server --list-modules dumps core on OpenBSD #2422

paultheweird commented Apr 10, 2015

fobser commented Apr 11, 2015

cmouse commented Apr 11, 2015

fobser commented Apr 11, 2015

cmouse commented Apr 11, 2015

fobser commented Apr 12, 2015

cmouse commented Apr 12, 2015

paultheweird commented Apr 14, 2015

cmouse commented Apr 14, 2015

cmouse commented Apr 14, 2015

paultheweird commented Apr 14, 2015

cmouse commented Apr 14, 2015

paultheweird commented Apr 14, 2015

cmouse commented Apr 14, 2015

paultheweird commented Apr 14, 2015

paultheweird commented Apr 14, 2015

cmouse commented Apr 14, 2015

paultheweird commented Apr 14, 2015

cmouse commented Apr 14, 2015

paultheweird commented Apr 14, 2015

cmouse commented Apr 14, 2015

paultheweird commented Apr 14, 2015

cmouse commented Apr 14, 2015

fobser commented Apr 14, 2015

cmouse commented Apr 14, 2015

paultheweird commented Apr 15, 2015

cmouse commented Apr 15, 2015

paultheweird commented Apr 15, 2015

rubenk commented May 5, 2015

ahupowerdns commented May 5, 2015

rubenk commented May 5, 2015

ahupowerdns commented May 5, 2015 via email

rubenk commented May 5, 2015

Habbie commented Jun 16, 2015

pdns_server --list-modules dumps core on OpenBSD #2422

pdns_server --list-modules dumps core on OpenBSD #2422

Comments

paultheweird commented Apr 10, 2015

fobser commented Apr 11, 2015

cmouse commented Apr 11, 2015

fobser commented Apr 11, 2015

cmouse commented Apr 11, 2015

fobser commented Apr 12, 2015

cmouse commented Apr 12, 2015

paultheweird commented Apr 14, 2015

cmouse commented Apr 14, 2015

cmouse commented Apr 14, 2015

paultheweird commented Apr 14, 2015

cmouse commented Apr 14, 2015

paultheweird commented Apr 14, 2015

cmouse commented Apr 14, 2015

paultheweird commented Apr 14, 2015

paultheweird commented Apr 14, 2015

cmouse commented Apr 14, 2015

paultheweird commented Apr 14, 2015

cmouse commented Apr 14, 2015

paultheweird commented Apr 14, 2015

cmouse commented Apr 14, 2015

paultheweird commented Apr 14, 2015

cmouse commented Apr 14, 2015

fobser commented Apr 14, 2015

cmouse commented Apr 14, 2015

paultheweird commented Apr 15, 2015

cmouse commented Apr 15, 2015

paultheweird commented Apr 15, 2015

rubenk commented May 5, 2015

ahupowerdns commented May 5, 2015

rubenk commented May 5, 2015

ahupowerdns commented May 5, 2015 via email

rubenk commented May 5, 2015

Habbie commented Jun 16, 2015