Suddenly can only get relay connection (server problem?) [ Edited ]

[Morac]

Edit: Found the problem. The drive isn't sending the external ip address in it's requests to wd2go.com as such the requests are invalid. No automated fix, but there is a manual workaround. Basically it's a bug in an executable. See last post for details.

I set up my My Book Live about 5 weeks ago.  At the time I set up port forwarding on my router and set up manual ports on the drive and got a "connected" status telling me port forwarding was working.  That worked fine until last night when I ended up disconnectnig the network cable from the My Book Live.  Since then I can't get the WD2Go service into direct connection (port forwarding) mode, it always goes to relay mode.  I didn't change anything on my router.  The only thing I did on the drive itself is enable SSH and change the root password.

I set up logging on my router and when I enable (or disable) remote access, the drive makes 4 or 5 connections to https://wd2go.com (198.107.148.110) and then simply connects to the relay server.  There are no incoming connection attempts at all.  I tested logging by connecting to the remote port I set up on my router and that shows up in my log (and I get a 403 Forbidden page in my browser), so logging is working.

The problem appears to be that the WD2Go service never even bothers to test to see if the ports I added are being forwarded.  It just switched to relay mode.

A few other things I've seen:

1. Sometimes the status temporarily changes to "failed", even though the relay connection is still up.

2. Sometimes the status temporarily changes to "connection" (not relay), even though I can see the relay connection being used.  Sometimes it looks like it really does switch to port forwarding, but then simply opens another relay connection.  Whatever it does, it switches back to relay pretty quickly (within a minute).

I don't want to use relay mode since it's slower and it's "leaking bytes".  I tracked the relay connection while it was not in use (idle) for 10 minutes and there was a total of 114.6 KB going back and forth since the drive periodically contacts wd2go.com and sometimes switches relay ip addresses for some reason.  That's 687.6 KB per hour, 16.12 MB per day and about 483 MB per month when not even using the connection.   That's rediculous that the My Book Live is now using nearly 500 MB of data a month when idle when it used to use nothing.  Especially since I have a capped service Internet service .

Direct mode worked fine for 50 days, how can I get the servers to switch back to direct mode?   I've sent an email to WD, but don't expect a useful response).

[WDTony]

Try disabling remote access and enabling it again.

[Morac]

Thanks for the response, but I've tried that.   I've also tried rebooting the drive and turning the drive off and leaving it off for about 10 minutes and plugging it in.

The problem appears to be on the server side.  The server does not appear to be checking the forwarded ports according to my router logs.  Actually it's not checking any ports since my logs show no incoming hits (blocked or accepted) at the time when the drive is "connecting".  I only see outgoing connections to port 443 on wd2go.com and a UDP connect to the relay server (ip varies).

The only thing I've changed on the drive since setting up remote access initialily is enabling SSH and changing the root password.  That wouldn't cause this would it?

[Morac]

I just tried something to test things.  In the past when I went to wd2go.com it would take me to mybooklive.device#####.wd2go.com : Port where ###### was my device id and port was the port number I chose to forward to 443.   After this stopped working the wd2go site stopped working and I had to relink it with my drive. .

I just went to wd2go and clicked on my drive and simply replaced mybooklive.device#####.wd2go.com with my home ip address : Port.  I got a login page and was able to log into my drive.  

So my drive is accessible via the Internet on the port I set up, but for whatever reason WD2Go is refusing to set up the dynamic DNS for it since it thinks the ports aren't accessible when they are.  Like I mentioned, there are no connection attempts on those ports so WD2Go isn't even testing if they are active or if it is, it's not testing my ip address.

[Morac]

Argh, I thought I had it working since I had turned off remote access while doing all the tests above and when I turned it back on, it showed a direct connection and there was no outbound connection to the relay, but about 2 minutes later the drive made a bunch of connections to wd2go.com and switched back to relay mode. 

I have no idea what's going on here only that my drive is accessible from the Internet despite what the WD2Go service thinks.

[Morac]

I think the DynDNS server might be screwed up.  My Book Live is showing that it's directly connected again and if I go to wd2go.com and log in and click my drive it tries to connect to https://mybooklive.device#####.wd2go.comORT/  where PORT is my 443 port forward port, but the it doesn't connect.  The reason being that mybooklive.device#####.wd2go.com is still pointing to the relay ip address, not my home address.  

It then switched back to relay mode and then the mybooklive.device#####.wd2go.com changed to the new relay ip address.

I've also noticed that when I disable sharing the mybooklive.device#####.wd2go.com address takes a long time to switch to unresolved even though the TTL on the nameserver is 10 seconds.  For example, I disabled remote access over a minute ago, but mybooklive.device#####.wd2go.com is still resolving.

So when it's set up to do a direct connection, it's not updating the mybooklive.device#####.wd2go.com to my ip address.  That could be why it's failing when trying to port check.  Though that doesn't explain why it periodicaly switches to direct mode and then back to relay.

I'm convinced this is a server problem, which means I can't fix it.

[Morac]

Just some more info, I noticed that status keeps changing from relay to portforwarding to relay periodically.  I was sshing into the drive at that time and dumped out the dynamincconfig.ini over the course of a few minutes.  It seems to switch to port forwarding mode for a few minutes and then switch back.  The entire time, my router reports the relay connection is still active (it's UDP though so it's hard to tell).  When it switches back to relay, it opens a new relay connection.

Basically it starts off with no external ip or external port, but it does have an external SSL port.   It gains the external port and switches to port forwarding (still no external ip address though).  It then somehow loses the exernal port and then switches back to relay mode.  It then stays like this a while and repeats the process.  It's been doing this quite often today.

The entire time it shows "portforwarded" the ip address for mybooklive.device######.wd2go.com still shows the relay ip address.  There's also a mybooklive-device######.wd2go.com which sometimes is the same ip address and sometimes not.  Some times it changes to "failed" instead of "relay" and the ip address associated with the dynamic dns goes away, but will change back to "relay" and get a relayed ip address.  It still won't actually use my ip address.   I have no idea why since I can't find anything in the log file about that.  I can't even figure out how it's possible to not know my external ip address since the drive is constantly contacting wd2go.com and that connection will have my ip address.

I have no idea what this means exactly but it doesn't look right to me.   I think it's caues the book to not go back to sleep as well.

Can anyone that has port forwarding working tell me if the "EXTERNAL_IP" field is supposed to be blank?

MyBookLive:/tmp# cat dynamicconfig.ini
SUBDOMAIN=""
DEVICEID="[censored]"
DEVICEAUTH="[censored]"
EXTERNAL_IP=""
EXTERNAL_PORT=""
EXTERNAL_SSL_PORT="443"
INTERNAL_IP="192.168.1.9"
INTERNAL_PORT="80"
DEVICE_SSL_PORT="443"
REMOTEACCESS="TRUE"
COMMUNICATION_STATUS="relayed"
DEFAULT_PORTS_ONLY="FALSE"
MANUAL_PORT_FORWARD="TRUE"
MANUAL_EXTERNAL_ROUTER_IP=""
MANUAL_EXTERNAL_HTTP_PORT="80"
MANUAL_EXTERNAL_HTTPS_PORT="443"
TOTAL_SETTINGS="17"

MyBookLive:/tmp# cat dynamicconfig.ini
SUBDOMAIN=""
DEVICEID="[censored]"
DEVICEAUTH="[censored]"
EXTERNAL_IP=""
EXTERNAL_PORT="80"
EXTERNAL_SSL_PORT="443"
INTERNAL_IP="192.168.1.9"
INTERNAL_PORT="80"
DEVICE_SSL_PORT="443"
REMOTEACCESS="TRUE"
COMMUNICATION_STATUS="portforwarded"
DEFAULT_PORTS_ONLY="FALSE"
MANUAL_PORT_FORWARD="TRUE"
MANUAL_EXTERNAL_ROUTER_IP=""
MANUAL_EXTERNAL_HTTP_PORT="80"
MANUAL_EXTERNAL_HTTPS_PORT="443"
TOTAL_SETTINGS="17"


MyBookLive:/tmp# cat dynamicconfig.ini
SUBDOMAIN=""
DEVICEID="[censored]"
DEVICEAUTH="[censored]"
EXTERNAL_IP=""
EXTERNAL_PORT=""
EXTERNAL_SSL_PORT="443"
INTERNAL_IP="192.168.1.9"
INTERNAL_PORT="80"
DEVICE_SSL_PORT="443"
REMOTEACCESS="TRUE"
COMMUNICATION_STATUS="portforwarded"
DEFAULT_PORTS_ONLY="FALSE"
MANUAL_PORT_FORWARD="TRUE"
MANUAL_EXTERNAL_ROUTER_IP=""
MANUAL_EXTERNAL_HTTP_PORT="80"
MANUAL_EXTERNAL_HTTPS_PORT="443"
TOTAL_SETTINGS="17"

MyBookLive:/tmp# cat dynamicconfig.ini
SUBDOMAIN=""
DEVICEID="[censored]"
DEVICEAUTH="[censored]"
EXTERNAL_IP=""
EXTERNAL_PORT=""
EXTERNAL_SSL_PORT="443"
INTERNAL_IP="192.168.1.9"
INTERNAL_PORT="80"
DEVICE_SSL_PORT="443"
REMOTEACCESS="TRUE"
COMMUNICATION_STATUS="relayed"
DEFAULT_PORTS_ONLY="FALSE"
MANUAL_PORT_FORWARD="TRUE"
MANUAL_EXTERNAL_ROUTER_IP=""
MANUAL_EXTERNAL_HTTP_PORT="80"
MANUAL_EXTERNAL_HTTPS_PORT="443"
TOTAL_SETTINGS="17"

edit: now it changed to failed:

MyBookLive:/tmp# cat dynamicconfig.ini
SUBDOMAIN=""
DEVICEID="[censored]"
DEVICEAUTH="[censored]"
EXTERNAL_IP=""
EXTERNAL_PORT=""
EXTERNAL_SSL_PORT="443"
INTERNAL_IP="192.168.1.9"
INTERNAL_PORT="80"
DEVICE_SSL_PORT="443"
REMOTEACCESS="TRUE"
COMMUNICATION_STATUS="failed"
DEFAULT_PORTS_ONLY="FALSE"
MANUAL_PORT_FORWARD="TRUE"
MANUAL_EXTERNAL_ROUTER_IP=""
MANUAL_EXTERNAL_HTTP_PORT="80"
MANUAL_EXTERNAL_HTTPS_PORT="443"
TOTAL_SETTINGS="17"

[Morac]

I guess this will be my last post unless I can get some answers.  I found out that it's possible to log the drive's connection to wd2go.com (the /usr/orion/communicationmanager/communicationmanagerd has a logtofile parameter).  I noticed two things when I stopped the communication manager and re-ran it with logging enabled.

1. When starting commnucationmanager reports two errors (don't know if they are important or not, but they could explain issue #2):

getaddrinfo() error : Name or service not known
getaddrinfo() error : Name or service not known

2. The command that communicationmanager is using to test to see if port forwarding works is missing the WAN ip address is the command:

https://www.wd2go.com/api/1.0/rest/device_communication/######?format=xml&external_ip=&external_port...   (first number is device id and second is the authorization, PORT is the 80 port).

If you notice the command doesn't say what external_ip to actual talk to.  That's very odd because the log file has the external ip address listed in it.  Anyway using the command above results in a 400 HTTP error with a description of " The request sent by the client was syntactically incorrect ()."

I wonder what happens if I actually put the correct external ip address in the command?  Let's see:

<device_communication><status>Success</status><req_device_id>####</req_device_id><resp_device_id>#####</resp_device_id><external_ip>#######</external_ip><remote_external_ip>null</remote_external_ip></device_communication>

Well what do you know it worked!   So basically the problem is that the drive is not sending the external ip address along with the request to check if port forwarding is working.

That seems like a bug in the /usr/orion/communicationmanager/communicationmanager program or something that's editing the configuration file (dynamicconfig.ini).

What I can't figure out is how to get the drive to actually send my external ip address in the request.  Any idea on how to fix this?

Edit:

As a work around I pretty much stopped remote access, then started it and quickly stopped the communicationmanager program before it opened a relay connection.  I then manually ran through the steps that the program would do, putting in my external_ip address.  That worked and now I've got a "direct" connection and port forwarding works remotely.

The problem with this set up is that since the communicationmanager program isn't running, any changes to my external ip won't get picked up and obvioulsy if the drive reboots that won't get picked up either.  Also there's likely some kind of server time out which will kick in at some point and remove the dynamic address and  mapping.  So it's really a poor substute at best.

I need to figure out what changed since when I originally set up port forwarding (in this version of the firmware) it worked.  Now it's not sending the external_ip address which prevents direct connection from working if communicationmanager program is running.

[Brondt]

Hi

I have the exact same issue, I think it started after upgrading the firmware April 26. 2012. The former version performed flawless, I have tried to see if I can revert to that version, but after what I've read it would be a bad thing to do, reverting that is.

Hope some one at WD will look in to it, as it is now the wd2go is kind of useless

Regards M.Brondt

[Brondt]

@Morac wrote:

"As a work around I pretty much stopped remote access, then started it and quickly stopped the communicationmanager program before it opened a relay connection."

As a ssh novice, how do you perform this, I know how to start the ssh session, but don't know how I put communicationmanager to a stop, can you help me?

Thanks

Brondt