T32 pause in streaming audio and video

[gearvault]

Hello,

I found out that T32 got rid of the 2 second pause at first access on my Ultra 6 but introduced a new pause some 5 minutes down the line which
halts an audio stream to my squeezebox and halts a video stream to my AC Ryan media player.
The pause varies in length, sometimes the player looses the stream and cannot reconnect and sometimes it picks up from
where is paused.
It seems that the pause varies about 5 minutes into the stream.
I am now back at 4.2.19 and the pause has disappeared.
(for your reference: I started using the betas because Logitech fiddled with the shares and rights
in the Squeezebox server software but they have solved that issue)

[btaroli]

This may coincide with a behavior I noticed with T32, which was elevated packet loss. I was noticing 1-2.5% with just a regular ping running over a couple hours during mixed use.

[gearvault]

Hi,

underneath the content of network_settings.log of another ultra 6 which is still running T32:
eth0 Link encap:Ethernet HWaddr 00:1F:33:EA:A3:00
inet addr:192.168.1.42 Bcast:192.168.1.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:9000 Metric:1
RX packets:32499 errors:0 dropped:1386 overruns:0 frame:0
TX packets:24820 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:6322163 (6.0 MiB) TX bytes:4805611 (4.5 MiB)
Interrupt:43 Memory:feadc000-0

eth1 Link encap:Ethernet HWaddr 00:1F:33:EA:A3:01
inet addr:192.168.1.43 Bcast:192.168.1.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:9000 Metric:1
RX packets:20918 errors:0 dropped:1386 overruns:0 frame:0
TX packets:1301 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:4294311 (4.0 MiB) TX bytes:389429 (380.3 KiB)
Interrupt:44 Memory:febdc000-0

Notice the number of dropped received packets. Is this due to a buffer that is too small?

On the recently downgraded ultra 6 (now 4.2.19) the packet loss is much less, percentage wise.
edit: Roughly same number of packets lost but 10 times more traffic.
So this confirms btaroli's findings!

cheers!