Bug 15724 - Errors when switching between music libraries
: Errors when switching between music libraries
Status: RESOLVED DUPLICATE of bug 15633
Product: SqueezePlay
Classification: Unclassified
Component: SB Server
: 7.5.x
: PC Other
: P1 normal (vote)
: 7.5.0
Assigned To: Felix Mueller
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2010-02-17 01:24 UTC by Felix Mueller
Modified: 2010-02-25 09:50 UTC (History)
8 users (show)

See Also:
Category: ---


Attachments
event id=nil issue when switching between libraries (30.07 KB, text/plain)
2010-02-17 01:25 UTC, Felix Mueller
Details
Disconnect and immediate reconnect to idle server (2.32 KB, text/plain)
2010-02-17 01:34 UTC, Felix Mueller
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Felix Mueller 2010-02-17 01:24:15 UTC
Some errors I see switching between TinySC, BigSC and test.mysb.com

Note: In this cases the network was always available.
Comment 1 Felix Mueller 2010-02-17 01:25:35 UTC
Created attachment 6537 [details]
event id=nil issue when switching between libraries

Some event id=nil Comet issues I do not understand at all.

Note: I did not even disconnect the network in this case. I only switched
between libraries (BigSC, TinySC and mysb.com).

Highlights from the log regarding datestatus subscription:

- 11:16:35 - subscribe to mysb.com (I think) with a valid event id = 21
- 11:16:35 - unsubscribed from Touch, event id = 15
- 11:16:35 - subscribe to mysb.com, event id = 21
- 11:16:37 - response from mysb.com failed due to invalid event id

I fail to understand why the datestatus event id could be invalid in this case.

7.5.0 r8504
Comment 2 Felix Mueller 2010-02-17 01:34:52 UTC
Created attachment 6538 [details]
Disconnect and immediate reconnect to idle server

- 10:17:22 - disconnect from idle sever (i.e. the one we switched away from)
- 10:17:23 - reconnect to this same server (I think unnecessarily)
- 10:18:23 - finally disconnected for good
Comment 3 Felix Mueller 2010-02-17 01:45:49 UTC
It is my understanding that the library switching mechanism temporarily connects to all available music sources to gather information about what servers are available and what players are connected to this servers.

So after a selection for a new music source has been made the connections to the not selected servers eventually time out.

But for some unknown reason sometimes the timed out connection disconnects and immediately reconnects only to time out again a bit later.

In my opinion this reconnection causes some issues we are seeing when switching between music libraries as this reconnections also initiate connection notifications to be sent out and in different places there is code acting upon this notifications.

Related bugs: 15627 and 15633
Comment 4 Felix Mueller 2010-02-17 01:46:48 UTC
Related bugs: Bug 15627 and bug 15633
Comment 5 Chris Owens 2010-02-18 10:01:53 UTC
QA to work on better repro instructions.  Maybe post to forums asking for more info, as well.
Comment 6 Chris Owens 2010-02-18 10:02:42 UTC
Oops I forgot to remove bug_meeting from these bugs.
Comment 7 Felix Mueller 2010-02-18 10:07:49 UTC
Just to be clear. I am stuck on this one and hope Alan can shed some light on it.
Comment 8 Alan Young 2010-02-19 07:14:17 UTC
Involved parts: SlimDiscovery(Applet), SlimBrowse, SlimServer and SlimProto
Comment 9 Felix Mueller 2010-02-19 07:18:03 UTC
Maybe also: ChooseMusicSource(Applet)
Comment 10 Alan Young 2010-02-23 08:46:48 UTC
Felix, I think that it is possible for this to go wrong under some circumstances, when a server does not notice that a player has disconnected and so continues to report that it is connected in serverStatus responses. It is possible that this would only occur with mySB. This would cause the SP's idea of which server the player is connected to to flip-flop. I think that it is unlikely to occur in normal usage and also likely to sort itself out after a minute of so at the outside.
Comment 11 Chris Owens 2010-02-25 09:50:11 UTC

*** This bug has been marked as a duplicate of bug 15633 ***