Boxee

Boxee compatibility issues with routers (mostly Linksys)

Details

  • Severity:
    Critical
  • Version:
    0.9.20
  • Operating System:
    Windows 7, Windows Vista, Windows XP
  • Number of attachments:
    3

Description

When Boxee is open and running from any windows platform, Boxee crashes from the opening of the Boxee application or sometimes it will stay running for couple minutes.

But I have found out that this is only happening if I am plugged into Linksys router products. So this is a major issue with compatibilities with Boxee and Linksys products.
I have testing this on Linksys WRT54G, RV042, and RV082 Linksys products.
I also tried to DMZ my PC behind the firewall and it still crashes.
Im not sure if anyone else is having this problem, but I have not been able to find any related postings about this major issue.
Please look into this and fix.
  1. boxee.log
    21/Feb/10 11:26 PM
    147 kB
    Brian C.
  2. boxee-04-01-10.log
    01/Apr/10 10:19 AM
    18 kB
    Brian C.
  1. screenshot-1.jpg
    158 kB
    01/Apr/10 10:04 AM

Issue Links

Activity

Hide
Rafael Mizrahi [Boxee] added a comment - 21/Feb/10 5:57 AM
Thanks allot Brian for investigating this issue.
you are not the only one experiencing Linksys+Boxee issues.

Please enable the advanced settings log
see http://support.boxee.tv/forums/49599/entries/43963
Start Boxee, and after experiencing the problem, close Boxee and attach your log file to the bug.
Show
Rafael Mizrahi [Boxee] added a comment - 21/Feb/10 5:57 AM Thanks allot Brian for investigating this issue. you are not the only one experiencing Linksys+Boxee issues. Please enable the advanced settings log see http://support.boxee.tv/forums/49599/entries/43963 Start Boxee, and after experiencing the problem, close Boxee and attach your log file to the bug.
Hide
Danita Zanre added a comment - 01/Mar/10 11:30 AM
Per a thread on the boxee forums, I'm adding info here. Please note, however, that my issue is not with Windows, but with OS X (Snow Leopard) on a Mac Mini - exact same symptoms it seems though as on the Windows 7 report here.

Quote:
Originally Posted by dconti
2) Add the following info in an entry to the bug in jira (http://jira.boxee.tv/browse/BOXEE-4441):
-- Model of your router and the firmware revision
WRT54G v2 with firmware v4.21.1
-- If you are using SMB and, if so, if your HD is plugged directly into your router, or if its separate
complicated - see below :-)
-- What version of boxee and what OS you run boxee on
OS X 10.6.2 Boxee 0.9.20.10708


Quote:
Also, a general question (for this thread) - can anyone repro this problem by using a different media app to crawl their smb share? For example, if you add the share to WMP and let it crawl, does it create the same problem?
/Quote

So, just to be a bit more clear here, I'll outline my network (which while not totally unique, might be a bit more complex than some, because I run a business out of my home, and have multiple servers here).

1. Comcast's cable modem is attached to an Astaro Security Gateway, which of course is in turn attached to a gigabit switch.
2. Most of the server hardware is in the server closet in the basement. However, there is an ethernet cable running from the gigabit switch to a closet on the first floor where the wireless router is located - and granted, it's older - a WRT54G v2 with firmware v4.21.1, which is the latest firmware for this particular wireless router.
3. The WRT65G also has a 4 port 10/100 switch, and is then connected from there up through the wall to the second floor to the D-Link DES-1105 switch. That switch currently services a VOIP box for ViaTalk, a TiVO Series 3, a Slingbox and the Mac Mini in question here.

When Boxee causes my connection to die, I simply have to unplug the D-Link DES-1105 switch to get connectivity back. So, I'm not sure what relationship it has to the Linksys wireless router, since that doesn't have any problems during the process, but there is indeed a wireless router "in the mix" so to speak.

Now, our media server is an OES2 (SuSE Linux) server, which shares up the volumes via CIFS (which is of course SMB). All Macs in the house use CIFS to connect to the shares on this server. It is used frequently throughout the day by everyone. The types of applications that access these shares are iTunes, Windows Media Player (rarely), Roxio Toast Video Player (to play Tivo transfer files that are also on the network), MultiPlex, EyeTV, and now also Boxee and Plex. Boxee is the only application that causes the problem. All other applications can access files on the shares all day long with no problems. The only other app on the Mac that I think will "crawl" is Plex (I think??). On the Mac Mini, Boxee is indeed the only app that causes any troubles like this.

I've just checked out one of the few Windows PCs here to see what would happen if I connected to that share and let it crawl in WMP. Took about a minute, and no disconnection from the network at all.
Show
Danita Zanre added a comment - 01/Mar/10 11:30 AM Per a thread on the boxee forums, I'm adding info here. Please note, however, that my issue is not with Windows, but with OS X (Snow Leopard) on a Mac Mini - exact same symptoms it seems though as on the Windows 7 report here. Quote: Originally Posted by dconti 2) Add the following info in an entry to the bug in jira (http://jira.boxee.tv/browse/BOXEE-4441): -- Model of your router and the firmware revision WRT54G v2 with firmware v4.21.1 -- If you are using SMB and, if so, if your HD is plugged directly into your router, or if its separate complicated - see below :-) -- What version of boxee and what OS you run boxee on OS X 10.6.2 Boxee 0.9.20.10708 Quote: Also, a general question (for this thread) - can anyone repro this problem by using a different media app to crawl their smb share? For example, if you add the share to WMP and let it crawl, does it create the same problem? /Quote So, just to be a bit more clear here, I'll outline my network (which while not totally unique, might be a bit more complex than some, because I run a business out of my home, and have multiple servers here). 1. Comcast's cable modem is attached to an Astaro Security Gateway, which of course is in turn attached to a gigabit switch. 2. Most of the server hardware is in the server closet in the basement. However, there is an ethernet cable running from the gigabit switch to a closet on the first floor where the wireless router is located - and granted, it's older - a WRT54G v2 with firmware v4.21.1, which is the latest firmware for this particular wireless router. 3. The WRT65G also has a 4 port 10/100 switch, and is then connected from there up through the wall to the second floor to the D-Link DES-1105 switch. That switch currently services a VOIP box for ViaTalk, a TiVO Series 3, a Slingbox and the Mac Mini in question here. When Boxee causes my connection to die, I simply have to unplug the D-Link DES-1105 switch to get connectivity back. So, I'm not sure what relationship it has to the Linksys wireless router, since that doesn't have any problems during the process, but there is indeed a wireless router "in the mix" so to speak. Now, our media server is an OES2 (SuSE Linux) server, which shares up the volumes via CIFS (which is of course SMB). All Macs in the house use CIFS to connect to the shares on this server. It is used frequently throughout the day by everyone. The types of applications that access these shares are iTunes, Windows Media Player (rarely), Roxio Toast Video Player (to play Tivo transfer files that are also on the network), MultiPlex, EyeTV, and now also Boxee and Plex. Boxee is the only application that causes the problem. All other applications can access files on the shares all day long with no problems. The only other app on the Mac that I think will "crawl" is Plex (I think??). On the Mac Mini, Boxee is indeed the only app that causes any troubles like this. I've just checked out one of the few Windows PCs here to see what would happen if I connected to that share and let it crawl in WMP. Took about a minute, and no disconnection from the network at all.
Hide
Danita Zanre added a comment - 01/Mar/10 11:39 AM
Oh - and I must add, that while I was directed to this bug from the forums, on closer examination, I think perhaps my issues are not entirely the same. While Boxee seems to lock up, what actually happens is that the Mac Mini loses its network connection, and the switch it is attached to must be cycled.
Show
Danita Zanre added a comment - 01/Mar/10 11:39 AM Oh - and I must add, that while I was directed to this bug from the forums, on closer examination, I think perhaps my issues are not entirely the same. While Boxee seems to lock up, what actually happens is that the Mac Mini loses its network connection, and the switch it is attached to must be cycled.
Hide
Dan Conti [Boxee] added a comment - 02/Mar/10 12:40 PM
Please see thread here: http://forum.boxee.tv/showthread.php?t=16410

I suspect what is happening is the linksys routers do not like either A) the udp packet sent to the .0 address on the subnet, or b) the subnet-scan nature of the packets being flooded out in short succession. In the case of Danita i wonder if this manifests itself as the port being wedged/shut off.

I am guessing there is a good reason that the packets are unicast to each node instead of being broadcast on the local network (addressing home routing/topology issues maybe?). Perhaps it is possible to spread the packets over a longer time period and see if that helps? Also i am not sure the unicast to the .0 address is needed and may in fact be problematic, it'd be worth trying to remove that (or seeing if someone can manually outbound block this at their firewall and see if that addresses the problem).

Lastly, it may be possible to broadcast to .255 by default and only do the subnet scan in some sort of troubleshooting mode?
Show
Dan Conti [Boxee] added a comment - 02/Mar/10 12:40 PM Please see thread here: http://forum.boxee.tv/showthread.php?t=16410 I suspect what is happening is the linksys routers do not like either A) the udp packet sent to the .0 address on the subnet, or b) the subnet-scan nature of the packets being flooded out in short succession. In the case of Danita i wonder if this manifests itself as the port being wedged/shut off. I am guessing there is a good reason that the packets are unicast to each node instead of being broadcast on the local network (addressing home routing/topology issues maybe?). Perhaps it is possible to spread the packets over a longer time period and see if that helps? Also i am not sure the unicast to the .0 address is needed and may in fact be problematic, it'd be worth trying to remove that (or seeing if someone can manually outbound block this at their firewall and see if that addresses the problem). Lastly, it may be possible to broadcast to .255 by default and only do the subnet scan in some sort of troubleshooting mode?
Hide
Dan Conti [Boxee] added a comment - 02/Mar/10 1:35 PM
Another set of thoughts after looking at a wireshark capture. Not all of these relate to info from Danita above.

I didn't see the packet to .0 on macos. It may be specific to the topology in the thread linked previously or to W7.

On a sparse network, doing the directed UDP messages to each port will generate a ton of ARP traffic. ARP is broadcast and over home wireless routers, flooding broadcast wireless traffic is pretty disastrous:

* Each multicast/broadcast packet will cause the wifi network to negotiate down to the slowest speed client on the network. So if the AP has a client running at 5mbps (like a phone) then the entire network will drop to 5mbps while the ARP traffic is sent.

* Multicast/broadcast traffic can't be sent instantly across the network and has to be queued, but most home APs do not have very large buffers; so if you flood the traffic you want to broadcast to the AP, it will fill the queue and drop some of your broadcasts. In this case this will manifest itself as failing to discover SMB shares on the local network.

* In Danita's case above, the wireless behavior is orthogonal, but the ARP flood may be causing problems for the dlink switch causing it to wedge.

At least some of the problems seem like they would be addressed by pacing the outbound netbios traffic (which in turn will pace the ARP traffic). On my mac I see 253 ARP requests in 20ms or about 12.5 packets/ms. If it's critical to use this mapping technique then i would suggest pacing it to a packet every 3-5ms. This will still get you discovery within ~ 1 second, but will increase the reliability and should help with the router compat problems.
Show
Dan Conti [Boxee] added a comment - 02/Mar/10 1:35 PM Another set of thoughts after looking at a wireshark capture. Not all of these relate to info from Danita above. I didn't see the packet to .0 on macos. It may be specific to the topology in the thread linked previously or to W7. On a sparse network, doing the directed UDP messages to each port will generate a ton of ARP traffic. ARP is broadcast and over home wireless routers, flooding broadcast wireless traffic is pretty disastrous: * Each multicast/broadcast packet will cause the wifi network to negotiate down to the slowest speed client on the network. So if the AP has a client running at 5mbps (like a phone) then the entire network will drop to 5mbps while the ARP traffic is sent. * Multicast/broadcast traffic can't be sent instantly across the network and has to be queued, but most home APs do not have very large buffers; so if you flood the traffic you want to broadcast to the AP, it will fill the queue and drop some of your broadcasts. In this case this will manifest itself as failing to discover SMB shares on the local network. * In Danita's case above, the wireless behavior is orthogonal, but the ARP flood may be causing problems for the dlink switch causing it to wedge. At least some of the problems seem like they would be addressed by pacing the outbound netbios traffic (which in turn will pace the ARP traffic). On my mac I see 253 ARP requests in 20ms or about 12.5 packets/ms. If it's critical to use this mapping technique then i would suggest pacing it to a packet every 3-5ms. This will still get you discovery within ~ 1 second, but will increase the reliability and should help with the router compat problems.
Hide
David Gilbert added a comment - 02/Mar/10 2:56 PM
I have a Sagem f@st 2404 router which also shows this problem which can only be resolved by rebooting the router. The router never causes any problems and remains running for months at a time without ever needing a reboot. That is until I started to use Boxee Beta on my ATV. Starting Boxee can often be the trigger that kills the routers wireless signal/performance.

The router firmware is;


3.24a4

Hoping this gets fixed soon as it puts me off using Boxee now.
Show
David Gilbert added a comment - 02/Mar/10 2:56 PM I have a Sagem f@st 2404 router which also shows this problem which can only be resolved by rebooting the router. The router never causes any problems and remains running for months at a time without ever needing a reboot. That is until I started to use Boxee Beta on my ATV. Starting Boxee can often be the trigger that kills the routers wireless signal/performance. The router firmware is; 3.24a4 Hoping this gets fixed soon as it puts me off using Boxee now.
Hide
Rafael Mizrahi [Boxee] added a comment - 18/Mar/10 5:08 AM
assigned to Amit and Dan to fix the nbtscan.
Show
Rafael Mizrahi [Boxee] added a comment - 18/Mar/10 5:08 AM assigned to Amit and Dan to fix the nbtscan.
Hide
Brian C. added a comment - 01/Apr/10 10:03 AM - edited
I downloaded and tried the link I recieved from Rafael "win32: boxee-0.9.21.11269.exe" and still same problems. I tried to create another log for latest build by going to Settings > Debug > Debug Level = Debug. But cant get into settings before it crashes. I attached a screen shot of error im getting, it doesnt saying anything that could help but here it is. Any other information needed please let me know what I can do to help fix this issue.
Show
Brian C. added a comment - 01/Apr/10 10:03 AM - edited I downloaded and tried the link I recieved from Rafael "win32: boxee-0.9.21.11269.exe" and still same problems. I tried to create another log for latest build by going to Settings > Debug > Debug Level = Debug. But cant get into settings before it crashes. I attached a screen shot of error im getting, it doesnt saying anything that could help but here it is. Any other information needed please let me know what I can do to help fix this issue.
Hide
Rafael Mizrahi [Boxee] added a comment - 01/Apr/10 10:08 AM
you can enable the advanced settings log
see http://support.boxee.tv/forums/49599/entries/43963
Show
Rafael Mizrahi [Boxee] added a comment - 01/Apr/10 10:08 AM you can enable the advanced settings log see http://support.boxee.tv/forums/49599/entries/43963
Hide
Dan Conti [Boxee] added a comment - 25/Apr/10 2:04 PM
Sorry, conflated two issues earlier here, this bug is specific to the boxee app crashing.

Brian, can you try disabling UPnP on your linksys products? And/or edit your guisettings.xml and change the 'client' value under <upnp> from true to false.

My guess is that we issue a search and the linksys router gives back some sort of malformed UPnP packet that causes the crash.

Failing this we can try a few other things.

Thanks for your help in tracking this down, very much appreciated.
Show
Dan Conti [Boxee] added a comment - 25/Apr/10 2:04 PM Sorry, conflated two issues earlier here, this bug is specific to the boxee app crashing. Brian, can you try disabling UPnP on your linksys products? And/or edit your guisettings.xml and change the 'client' value under <upnp> from true to false. My guess is that we issue a search and the linksys router gives back some sort of malformed UPnP packet that causes the crash. Failing this we can try a few other things. Thanks for your help in tracking this down, very much appreciated.
Hide
Rafael Mizrahi [Boxee] added a comment - 21/Jun/10 10:36 AM
Beau Brooks, reported at issue BOXEE-4035:
"This seems to be resolved for me after the OSX 10.6.4 upgrade and applying boxee 0.9.21.12089"
Show
Rafael Mizrahi [Boxee] added a comment - 21/Jun/10 10:36 AM Beau Brooks, reported at issue BOXEE-4035: "This seems to be resolved for me after the OSX 10.6.4 upgrade and applying boxee 0.9.21.12089"
Hide
Rafael Mizrahi [Boxee] added a comment - 06/Sep/10 8:40 AM
Brian, Danita,
did you managed to workaround the issue by disabling upnp on your Linksys router or on the Boxee client?
Show
Rafael Mizrahi [Boxee] added a comment - 06/Sep/10 8:40 AM Brian, Danita, did you managed to workaround the issue by disabling upnp on your Linksys router or on the Boxee client?
Hide
Brian C. added a comment - 06/Sep/10 11:03 AM
I am now having the same issue but this time I'm not using the Linksys RV042 router. I am using a Sonicwall TZ 180 Wireless, this issue still has not been resolved.
Show
Brian C. added a comment - 06/Sep/10 11:03 AM I am now having the same issue but this time I'm not using the Linksys RV042 router. I am using a Sonicwall TZ 180 Wireless, this issue still has not been resolved.
Hide
Brian C. added a comment - 06/Sep/10 11:05 AM
Do I need to have some $10,000 Cisco router to use Boxee? Im begining to wonder when the Boxee Box comes out will it do that same thing with my equipment.
Show
Brian C. added a comment - 06/Sep/10 11:05 AM Do I need to have some $10,000 Cisco router to use Boxee? Im begining to wonder when the Boxee Box comes out will it do that same thing with my equipment.
Hide
Dan Conti [Boxee] added a comment - 06/Sep/10 12:39 PM
Hi Brian -

Do you have any routers that you can use successfully? Or do all of your routers, linksys or not, result in a crash?

Have you tried disabling the upnp client as suggested previously? This will help us narrow down the issue.

Is there anything else about your home network topology that we should know? Are you using proxies, do you have NAS devices or UPNP servers that are being indexed, etc?

Please understand that issues like this, where it repros in just your environment, can be difficult to diagnose and may take a bit of "back and forth" before we can narrow it down sufficiently to address.

Thanks,
-Dan
Show
Dan Conti [Boxee] added a comment - 06/Sep/10 12:39 PM Hi Brian - Do you have any routers that you can use successfully? Or do all of your routers, linksys or not, result in a crash? Have you tried disabling the upnp client as suggested previously? This will help us narrow down the issue. Is there anything else about your home network topology that we should know? Are you using proxies, do you have NAS devices or UPNP servers that are being indexed, etc? Please understand that issues like this, where it repros in just your environment, can be difficult to diagnose and may take a bit of "back and forth" before we can narrow it down sufficiently to address. Thanks, -Dan
Hide
Rafael Mizrahi [Boxee] added a comment - 19/Jul/11 10:42 AM
Issue closed as obsolete.
We will be releasing a new PC/Mac/Ubuntu in the fall.
Please test with this new version and if this is still occurring please feel free to open a new ticket.
Show
Rafael Mizrahi [Boxee] added a comment - 19/Jul/11 10:42 AM Issue closed as obsolete. We will be releasing a new PC/Mac/Ubuntu in the fall. Please test with this new version and if this is still occurring please feel free to open a new ticket.

People

Vote (1)
Watch (3)

Dates

  • Created:
    20/Feb/10 12:10 PM
    Updated:
    19/Jul/11 10:42 AM
    Resolved:
    19/Jul/11 10:42 AM