Bug 14570 - Apple Dual-Band Airport seemingly incompatible with SqueezePlay-based Squeezeboxes
: Apple Dual-Band Airport seemingly incompatible with SqueezePlay-based Squeeze...
Status: NEW
Product: Logitech Media Server
Classification: Unclassified
Component: Sync
: unspecified
: PC Other
: -- normal with 2 votes (vote)
: ---
Assigned To: Unassigned bug - please assign me!
http://forums.slimdevices.com/showthr...
:
Depends on:
Blocks: 13092
  Show dependency treegraph
 
Reported: 2009-10-04 12:55 UTC by Matt Wise
Modified: 2011-11-06 23:23 UTC (History)
9 users (show)

See Also:
Category: Bug


Attachments
server.log of the failing sync (7.39 MB, text/plain)
2009-10-04 12:56 UTC, Matt Wise
Details
Bandwidth graph from Airports (74.01 KB, image/png)
2009-10-04 12:58 UTC, Matt Wise
Details
Last 1000 lines of messages file from misbehaving baby (84.02 KB, text/plain)
2009-10-04 13:05 UTC, Matt Wise
Details
7.4 official Server.log FLAC and Rhapsody sync errors (27.93 KB, text/plain)
2009-10-07 10:13 UTC, Nikolaj Christensen
Details
7.4.1-28791 still Rhapsody and FLAC sync errors (52.29 KB, text/plain)
2009-10-07 10:26 UTC, Nikolaj Christensen
Details
SqueezeboxServer-7.4.1-28791. MP3 sync not tight (17.98 KB, text/plain)
2009-10-08 08:53 UTC, Nikolaj Christensen
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Matt Wise 2009-10-04 12:55:30 UTC
In the saga of ongoing sync issues, heres a log file from syncing that is failing miserably this morning. I have two players, both are on separate Airport Extreme's (graph attached), each running on 2.4GHz. Both players Synced playing Pandora worked great for ~4 hours. 

I started playing FLAC albums... everything seemed OK. At one point I had to reboot an airport for a configuration change. When it came back up, nothing worked after that. The songs keep restarting at the beginning, and the Boom is playing the audio but the baby just keeps 're-buffering'. Playback never really starts up. Skipping the song did not solve the issue either. In order to solve the issue I had to physically reboot the Baby. When it came back up, playback started working again.
Comment 1 Matt Wise 2009-10-04 12:56:46 UTC
Created attachment 6009 [details]
server.log of the failing sync
Comment 2 Matt Wise 2009-10-04 12:58:39 UTC
Created attachment 6010 [details]
Bandwidth graph from Airports

You can see here that at around 12:05 or 12:10 I started the playback. The playback was fine for about 10 minutes before the playback problems began and the bandwidth drops to near zero. I rebooted the Baby somewhere around 12:25-12:30 and then playback begins to work again. 

These are the only two devices (other than two jives) on the 2.4GHz spectrum in my house...
Comment 3 Matt Wise 2009-10-04 13:04:40 UTC
The rebuffering problem just started again. There were no wifi interruptions this time, it just happened on its own and will not resolve itself. (other than perhaps rebooting the baby). 

Looking at the Baby itself you can see its CPU is maxed out. I can barely type on it via remote SSH. 

Mem: 52952K used, 9136K free, 0K shrd, 7312K buff, 16624K cached
CPU:  72% usr  27% sys   0% nic   0% idle   0% io   0% irq   0% sirq
Load average: 1.86 1.99 1.66 3/65 1188
  PID  PPID USER     STAT   VSZ %MEM %CPU COMMAND
  740     1 root     R    30036  48%  57% /usr/bin/jive 
 1187  1184 root     R     2692   4%  19% top 
  783   740 root     S     7108  11%  10% jive_alsa -d default -c dac -b 20000 -p 2 -s 16 -f 3 
  731     1 root     S     1892   3%   5% /usr/sbin/wpa_supplicant -B -Dwext -ieth1 -c/etc/wpa_supplicant.conf 

I've attached the messages file with the last 1000 lines in it or so. It's clearly having some kind of playback issue that I think is causing the sync issue actually.
Comment 4 Matt Wise 2009-10-04 13:05:08 UTC
Created attachment 6011 [details]
Last 1000 lines of messages file from misbehaving baby
Comment 5 Matt Wise 2009-10-04 13:09:02 UTC
After that 1000 line log I postd above, Baby rebooted on its own. Its come back up and is continuing to have audio underrun errors.
Comment 6 Matt Wise 2009-10-04 13:14:00 UTC
Ok... more info. I switched from FLAC to Pandora. The players are currently playing, but I am still seeing audio underrun errors. The sync is off by a good 2-3 seconds and staying that way right now.
Comment 7 James Richardson 2009-10-05 08:20:03 UTC
*** Bug 14567 has been marked as a duplicate of this bug. ***
Comment 8 James Richardson 2009-10-06 12:45:36 UTC
*** Bug 14626 has been marked as a duplicate of this bug. ***
Comment 9 James Richardson 2009-10-06 12:48:58 UTC
Is this a dupe or related to bug 13092?
Comment 10 Matt Wise 2009-10-06 13:32:27 UTC
Alan is pretty darn sure that the other bug is a network issue. I can easily pass the 5000KBPs stream test in my house with multiple players at the same time -- so its not a bandwidth issue.
Comment 11 Nikolaj Christensen 2009-10-07 10:13:17 UTC
Created attachment 6050 [details]
7.4 official Server.log FLAC and Rhapsody sync errors

My original bug got closed as a dupe, but it's not entirely the same as this bug, but anyway. Running 7.4 official receiver is 4-5 seconds ahead of boom when syncing FLAC or Rhapsody. Attaching server.log with player.info and player.sync debug info. Please let me know if you need anything else thanks.
Comment 12 Nikolaj Christensen 2009-10-07 10:26:24 UTC
Created attachment 6051 [details]
7.4.1-28791 still Rhapsody and FLAC sync errors

I upgraded to SqueezeboxServer-7.4.1-28791. Now it seems one receiver and one boom syncs ok with FLAC but still no Rhapsody sync. Adding another boom to the sync group breaks sync for both FLAC and Rhapsody.
Comment 13 Nikolaj Christensen 2009-10-08 08:53:16 UTC
Created attachment 6069 [details]
SqueezeboxServer-7.4.1-28791. MP3 sync not tight

Here is another server.log. This time it's an mp3 playing. Sync was kinda of there, but seemed somewhat off still. Less than 0.3 seconds though.

Does this log show any sync errors?
Comment 14 Alan Young 2009-10-09 01:46:56 UTC
Nikolaj, It looks like something is messing with your server clock. What do you use to set the time on your server?
Comment 15 Alan Young 2009-10-09 02:06:06 UTC
Matt, some definite weird stuff in your log. What are all the 

Slim::Player::SqueezePlay::updateCapabilities (128) unknown capability: ---=1, ignored

lines about? Can you capture a connect with network.protocol.slimproto=info enabled so we can see the capabilities being passed please?
Comment 16 Nikolaj Christensen 2009-10-09 08:03:40 UTC
(In reply to comment #14)
> Nikolaj, It looks like something is messing with your server clock. What do you
> use to set the time on your server?

It's a windows 7 ultimate server that currently uses "Internet time" synchronization.
Comment 17 Nikolaj Christensen 2009-10-09 08:15:03 UTC
- I did a resynchronize time which set the server back 1-2 minutes and restarted the server. Now Rhapsody, mp3, flac and internet radio seems to sync well between the two booms.

Anything I should do to avoid the server time causing problems? Using a fixed clock?
Comment 18 Matt Wise 2009-10-09 09:17:03 UTC
Alan,
  Andy says that error is normal -- it has to do with a line being commented out of the transcoder config file. I'll get a capture next time I can reproduce this with the logging you requested though. 

  For what its worth, I think that some of this might be solved with with the now playing screen memory leak bug... I was having serious playback issues this morning and it turned out I was out of memory.
Comment 19 Ben Klaas 2009-10-09 10:02:28 UTC
...not my bug to own
Comment 20 Andy Grundman 2009-10-09 10:11:53 UTC
The --- is the disabled ALAC codec.
Comment 21 Matt Wise 2009-10-09 17:49:42 UTC
Well I'll be damned... on a whim I decided to turn Wireless off on my (farther away) "server room" Airport (Gig-E, Single Band, N). When I did that, I actually nearly lost the entire ability to stream FLAC at all! I had a Boom (wireless, 20ft from dual-band airport, line of sight), a Baby (wireless, 20 ft from airport, one wall in the way), and a Touch (wired ethernet) synced... I couldn't get ANY music to play. Rebuffering all over the place, Boom even had such a hard time with the wireless that it crashed and rebooted! 

I switched things around... disabled WiFi on my Dual-Band GigE airport completely and turned WiFi back on the server room airport (which is farther away, through multiple walls)... almost instantly music started playing out of all 3 players! SYNCED! 

I think I might be crazy... but we'll see how it plays tonight. I suspect that there is something wrong either with my dual band airport, or the interaction between this airport and our products... the syncing is now damn near perfect, and all the players are very responsive! 

If this continues to work all weekend, I'll bring the airport into the office on Monday for further examination.
Comment 22 Alan Young 2009-10-09 19:03:28 UTC
Nikolaj, I'm not sure what to suggest as I'm not sure what options are available under Windows. Ideal would be an NTP client running continuously. The sync code does not really like small jumps backwards in server time (where 30ms < small < 50s).

The fix proposed in bug 14352 may be of help, but this is not yet committed for release.

Otherwise, restarting SC after the time adjustment would be one option, or running the time adjustment sufficiently often that all the adjustments are small.
Comment 23 Nikolaj Christensen 2009-10-10 08:29:24 UTC
- Thanks for the response Alan. I'll see what I can find regarding timesync. I do think that many users will have similar windows setups, so it would be great if the code could handle these time updates transparently.
Comment 24 Nikolaj Christensen 2009-10-10 08:43:51 UTC
- I have used this tutorial:

http://www.cumps.be/howto-set-up-ntp-on-windows-vista/

To set ntp updates to every 12 hours.

Was set to every 7 days as default. Let's see if that helps.
Comment 25 Matt Wise 2009-10-12 10:49:38 UTC
Just another comment from the forums that echoes the issues I have had:

http://forums.slimdevices.com/showpost.php?p=471116&postcount=7
"I had one and had problems with SBs, so I sent it back."
Comment 26 Matt Wise 2009-10-12 10:50:10 UTC
*** Bug 14368 has been marked as a duplicate of this bug. ***
Comment 27 Nikolaj Christensen 2009-10-12 13:44:18 UTC
- this morning players were out of sync again. Skimming the logs it shows that the ntp updated _after_ squeezebox server started after my daily scheduled reboot. Gonna try and schedule a manual ntp update before the reboot and disable automatic ntp updates.
Comment 28 Matt Wise 2009-10-19 19:08:49 UTC
Nikolaj,
  If you continue to have issues related to the NTP problem, could you please open a separate bug. It seems that this bug really has been broken down into an incompatibility with Apple Dual Band Airports. 

Felix,
 It seems that there are multiple reports of problems with the Dual Band airport. Do you expect this to be an issue you can solve remotely, or should we just send you the airport ASAP so that we can get a fix in for 7.4.x+? This is the only Airport Extreme that Apple ships now, so its likely to become a worse and worse issue as time goes on.
Comment 29 Andy Grundman 2009-10-19 19:11:22 UTC
FWIW, my dual-band AE has been working great.
Comment 30 Felix Mueller 2009-10-20 04:01:34 UTC
Matt: I talked to Richard about this and we both think it has something to do with having two APs running than one being dual band.

For now he suggests the following:

- Turn off all your players (i.e. take them off the wireless network)
- Only configure _one_ access point
- Turn on your players again (i.e. reconnect them to your network)
- Test if you still see the issue

- Repeat above with your other access point

Thanks
Felix
Comment 31 Chris Owens 2009-10-21 09:47:45 UTC
Moving these bugs to P4 to make room for moving P1.5 bugs to P2
Comment 32 Pat Ransil 2009-10-23 05:09:45 UTC
Administrative move of 7.5 bugs. All P2, P3, P4 being downgraded one level. Will then split P1s.
Comment 33 Chris Owens 2009-12-02 13:58:36 UTC
Matt, why is this bug assigned to you again?
Comment 34 Michael Herger 2009-12-04 00:29:41 UTC
This can't be a deal breaker.
Comment 35 Matt Wise 2009-12-04 08:17:50 UTC
Hah, i have no idea why this is assigned to me. That seems like a mistake.
Comment 36 Chris Owens 2010-03-08 11:17:25 UTC
Moving P3 and lower bugs to next release target
Comment 37 Alan Young 2011-11-06 23:23:46 UTC
Unassigned bugs cannot have a priority.