Re: openvpn in rdomain hangs

2014-02-11 Thread antoine 
Chris Cappuccio chris at nmedia.net writes:

 
 Giancarlo Razzolini [grazzolini at gmail.com] wrote:
  I've used rdomains, but not for this. In this case I would use mpath and
  pf only. I really do not see the need for using rdomains in this case.
  It introduces too much complexity for a simple thing.
  
 
 That's nice. But what about the problem?
 
 

Hi,

Did you find a solution  as I have the same issue ?

Thanks

Re: openvpn in rdomain hangs

2014-02-06 Thread Chris Cappuccio 
Giancarlo Razzolini [grazzol...@gmail.com] wrote:
 I've used rdomains, but not for this. In this case I would use mpath and
 pf only. I really do not see the need for using rdomains in this case.
 It introduces too much complexity for a simple thing.
 

That's nice. But what about the problem?

openvpn in rdomain hangs

2014-02-05 Thread Tomoyuki Sakurai 
hi misc,

i'm trying to run OpenBSD with two default gateways, one for openvpn and
another for everything else. openvpn is in rdomain 1 and everything works fine.

OpenBSD 5.5-beta (GENERIC.MP) #284: Mon Feb  3 07:57:32 MST 2014
t...@amd64.openbsd.org:/usr/src/sys/arch/amd64/compile/GENERIC.MP

# cat hostname.em1
!echo starting em1
rdomain 1
group vpn
inet yy.yy.yy.yy/28
!/sbin/route -T 1 add default yy.yy.yy.default
!echo -n starting sshd in rdomain 1
!route -T 1 exec /etc/rc.d/sshd start  echo .
!echo -n starting openvpn in rdomain 1
!install -d -o _openvpn -g _openvpn -m 0755 /var/run/openvpn 
/sbin/route -T 1 exec /usr/local/sbin/openvpn --daemon --config
/etc/openvpn/server.conf  echo .

# cat hostname.tun0
!echo starting tun0
up
10.100.16.1 10.100.16.2 netmask 0x
!route add 10.100.16.0/24 10.100.16.2

however, openvpn cannot be restarted in multi-user mode. it hangs in
get_default_gateway().

http://www.openssh.com/cgi-bin/cvsweb/ports/net/openvpn/patches/patch-src_openvpn_route_c?rev=1.2;content-type=text%2Fplain

ktrace shows that read(2) to routing socket does not return.

 10068 openvpn  CALL  socket(PF_ROUTE,SOCK_RAW,0)
 10068 openvpn  RET   socket 3
 10068 openvpn  CALL  sigprocmask(SIG_BLOCK,~0)
 10068 openvpn  RET   sigprocmask 0
 10068 openvpn  CALL  mprotect(0x10ee093000,0x2000,0x3PROT_READ|PROT_WRITE)
 10068 openvpn  RET   mprotect 0
 10068 openvpn  CALL  mprotect(0x10ee093000,0x2000,0x1PROT_READ)
 10068 openvpn  RET   mprotect 0
 10068 openvpn  CALL  sigprocmask(SIG_SETMASK,0)
 10068 openvpn  RET   sigprocmask ~0x10100SIGKILL|SIGSTOP
 10068 openvpn  CALL  write(0x3,0x10ee1949c0,0x80)
 10068 openvpn  GIO   fd 3 wrote 128 bytes
   
\M^@\0\^E\^D\0\0\0\0\0\0\0\0\^E\0\0\0\^C\0\0\0\0\0\0\0\0\0\0\0\^A\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\

\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\^P\^B\0\0\0\0\0\0\0\0\0\0\0\0\0\0\^P\^B\
\0\0\0\0\0\0\0\0\0\0\0\0\0\000
 10068 openvpn  RET   write 128/0x80
 10068 openvpn  CALL  sigprocmask(SIG_BLOCK,~0)
 10068 openvpn  RET   sigprocmask 0
 10068 openvpn  CALL  mprotect(0x10ee093000,0x2000,0x3PROT_READ|PROT_WRITE)
 10068 openvpn  RET   mprotect 0
 10068 openvpn  CALL  mprotect(0x10ee093000,0x2000,0x1PROT_READ)
 10068 openvpn  RET   mprotect 0
 10068 openvpn  CALL  sigprocmask(SIG_SETMASK,0)
 10068 openvpn  RET   sigprocmask ~0x10100SIGKILL|SIGSTOP
 10068 openvpn  CALL  read(0x3,0x10ee1949c0,0x260)

when invoked from hostname.em1 during boot, read(2) immediately returned with
ESRCH. according to route(4), messages written to the socket should be
returned. how can read(2) be blocked?
the behavior is same on 5.4, 5.3 and current.

-- 
Tomoyuki Sakurai

Re: openvpn in rdomain hangs

2014-02-05 Thread Giancarlo Razzolini 
Em 05-02-2014 07:09, Tomoyuki Sakurai escreveu:
 hi misc,

 i'm trying to run OpenBSD with two default gateways, one for openvpn and
 another for everything else. openvpn is in rdomain 1 and everything works 
 fine.

 OpenBSD 5.5-beta (GENERIC.MP) #284: Mon Feb  3 07:57:32 MST 2014
 t...@amd64.openbsd.org:/usr/src/sys/arch/amd64/compile/GENERIC.MP

 # cat hostname.em1
 !echo starting em1
 rdomain 1
 group vpn
 inet yy.yy.yy.yy/28
 !/sbin/route -T 1 add default yy.yy.yy.default
 !echo -n starting sshd in rdomain 1
 !route -T 1 exec /etc/rc.d/sshd start  echo .
 !echo -n starting openvpn in rdomain 1
 !install -d -o _openvpn -g _openvpn -m 0755 /var/run/openvpn 
 /sbin/route -T 1 exec /usr/local/sbin/openvpn --daemon --config
 /etc/openvpn/server.conf  echo .

 # cat hostname.tun0
 !echo starting tun0
 up
 10.100.16.1 10.100.16.2 netmask 0x
 !route add 10.100.16.0/24 10.100.16.2

 however, openvpn cannot be restarted in multi-user mode. it hangs in
 get_default_gateway().

 http://www.openssh.com/cgi-bin/cvsweb/ports/net/openvpn/patches/patch-src_openvpn_route_c?rev=1.2;content-type=text%2Fplain

 ktrace shows that read(2) to routing socket does not return.

  10068 openvpn  CALL  socket(PF_ROUTE,SOCK_RAW,0)
  10068 openvpn  RET   socket 3
  10068 openvpn  CALL  sigprocmask(SIG_BLOCK,~0)
  10068 openvpn  RET   sigprocmask 0
  10068 openvpn  CALL  mprotect(0x10ee093000,0x2000,0x3PROT_READ|PROT_WRITE)
  10068 openvpn  RET   mprotect 0
  10068 openvpn  CALL  mprotect(0x10ee093000,0x2000,0x1PROT_READ)
  10068 openvpn  RET   mprotect 0
  10068 openvpn  CALL  sigprocmask(SIG_SETMASK,0)
  10068 openvpn  RET   sigprocmask ~0x10100SIGKILL|SIGSTOP
  10068 openvpn  CALL  write(0x3,0x10ee1949c0,0x80)
  10068 openvpn  GIO   fd 3 wrote 128 bytes

 \M^@\0\^E\^D\0\0\0\0\0\0\0\0\^E\0\0\0\^C\0\0\0\0\0\0\0\0\0\0\0\^A\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\
 
 \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\^P\^B\0\0\0\0\0\0\0\0\0\0\0\0\0\0\^P\^B\
 \0\0\0\0\0\0\0\0\0\0\0\0\0\000
  10068 openvpn  RET   write 128/0x80
  10068 openvpn  CALL  sigprocmask(SIG_BLOCK,~0)
  10068 openvpn  RET   sigprocmask 0
  10068 openvpn  CALL  mprotect(0x10ee093000,0x2000,0x3PROT_READ|PROT_WRITE)
  10068 openvpn  RET   mprotect 0
  10068 openvpn  CALL  mprotect(0x10ee093000,0x2000,0x1PROT_READ)
  10068 openvpn  RET   mprotect 0
  10068 openvpn  CALL  sigprocmask(SIG_SETMASK,0)
  10068 openvpn  RET   sigprocmask ~0x10100SIGKILL|SIGSTOP
  10068 openvpn  CALL  read(0x3,0x10ee1949c0,0x260)

 when invoked from hostname.em1 during boot, read(2) immediately returned with
 ESRCH. according to route(4), messages written to the socket should be
 returned. how can read(2) be blocked?
 the behavior is same on 5.4, 5.3 and current.

I've used rdomains, but not for this. In this case I would use mpath and
pf only. I really do not see the need for using rdomains in this case.
It introduces too much complexity for a simple thing.

Cheers,

-- 
Giancarlo Razzolini
GPG: 4096R/77B981BC