[tor-bugs] #949 [Tor - Relay]: Tor server can't bind itself to a ip address and dies.

[Tor Bug Tracker & Wiki]

#949: Tor server can't bind itself to a ip address and dies.
--------------------------+-------------------------------------------------
  Reporter:  mavior       |       Owner:  weasel  
      Type:  defect       |      Status:  closed  
  Priority:  minor        |   Milestone:          
 Component:  Tor - Relay  |     Version:  0.2.0.34
Resolution:  invalid      |    Keywords:          
    Parent:               |  
--------------------------+-------------------------------------------------
Changes (by weasel):

  * status:  new => closed
  * resolution:  None => invalid


Old description:

> I am using Tor version 0.2.0.34 (r18423).
> I have set up Tor to bind the server to the localhost loopback and to a
> local network ip address.
> By default the daemon is configured to start in runlevel 2 with S20tor.
> When the system starts, there is a race condition of the daemons on the
> same run-level and if the kernel
> hasn't yet registered the ip address of the local network
> interface because it's querying in dhcp mode,
> tor puts a message on stdout on boot that it wasn't able to bind itself
> to the LAN address specified and dies.
>
> This is very annoying,because I often start the system and then I don't
> find the expected instance of tor running,
> even if the interface and the local ip address are working.
> For me it's better if the started instance of tor,keep polling for
> specified ip address instead of die.
>

> Also moving the name of the file to something like S90tor on all
> runlevels seems to help partially as the problem seems just to slightly
> decrease.(this confirm the race condition problem)
>
> I can't attach any log output as it does not output nothing relevant in
> the logs...it just dies.
>
> [Automatically added by flyspray2trac: Operating System: Other Linux]

New description:

 I am using Tor version 0.2.0.34 (r18423).
 I have set up Tor to bind the server to the localhost loopback and to a
 local network ip address.
 By default the daemon is configured to start in runlevel 2 with S20tor.
 When the system starts, there is a race condition of the daemons on the
 same run-level and if the kernel
 hasn't yet registered the ip address of the local network
 interface because it's querying in dhcp mode,
 tor puts a message on stdout on boot that it wasn't able to bind itself to
 the LAN address specified and dies.

 This is very annoying,because I often start the system and then I don't
 find the expected instance of tor running,
 even if the interface and the local ip address are working.
 For me it's better if the started instance of tor,keep polling for
 specified ip address instead of die.


 Also moving the name of the file to something like S90tor on all runlevels
 seems to help partially as the problem seems just to slightly
 decrease.(this confirm the race condition problem)

 I can't attach any log output as it does not output nothing relevant in
 the logs...it just dies.

 [Automatically added by flyspray2trac: Operating System: Other Linux]

--

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/949#comment:7>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online