[jitsi-dev] package loss 60% in jitsi-meet


#1

Hello,

I had lately a self-hosted jitsi-meet conference with 10 participants,
unfortunately many clients had package loss 50-60% seeing in status
pop-up (if you click on bars icon), I've looked at server logs after
that and only found something like this.

···

###################
2015-05-16 11:55:05.251 INFO: [45]
org.jitsi.impl.neomedia.MediaStreamImpl.info()
Receive stream stats: discarded RTP packets: 0
Receive stream stats: decoded with FEC: 0
2015-05-16 11:55:05.251 INFO: [45] net.sf.fmj.media.Log.info()
Stopping RTPSourceStream.
2015-05-16 11:55:05.257 SEVERE: [2828]
org.jitsi.impl.neomedia.transform.SinglePacketTransformer.error()
Failed to transform RawPacket(s)!
java.lang.NullPointerException
2015-05-16 11:55:05.259 INFO: [45]
org.jitsi.impl.neomedia.MediaStreamImpl.info() rtpstat:call stats for
outgoing audio stream SSRC: 785767589
rtpstat:bytes sent: 0
rtpstat:RTP sent: 0
rtpstat:remote reported min interarrival jitter: 0.0ms
rtpstat:remote reported max interarrival jitter: 18.666666666666668ms
rtpstat:local collisions: 0
rtpstat:remote collisions: 0
rtpstat:RTCP sent: 0
rtpstat:transmit failed: 0
2015-05-16 11:55:05.260 SEVERE: [2828]
org.jitsi.impl.neomedia.rtp.translator.OutputDataStreamImpl.error()
Failed to translate RTP packet
java.lang.NullPointerException
2015-05-16 11:55:05.260 INFO: [45]
org.jitsi.impl.neomedia.MediaStreamImpl.info() rtpstat:call stats for
incoming stream SSRC: 782004271
rtpstat:packets received: 327451
rtpstat:bytes received: 29828490
rtpstat:packets lost: 37
rtpstat:min interarrival jitter: 0
rtpstat:max interarrival jitter: 7865
rtpstat:RTCPs received: 36424
rtpstat:bad RTCP packets: 0
rtpstat:bad RTP packets: 157
rtpstat:local collisions: 0
rtpstat:malformed BYEs: 0
rtpstat:malformed RRs: 0
rtpstat:malformed SDESs: 0
rtpstat:malformed SRs: 0
rtpstat:packets looped: 0
rtpstat:remote collisions: 0
rtpstat:SRs received: 7930
rtpstat:transmit failed: 0
rtpstat:unknown types: 157
########################

I've checked munin graphs on server and doesn't found something
unusual (no errors on network interface either), all clients was
connected to 1GB LAN (no WiFi) with cpu last about 50-60%

Is there anyway do debug what caused it? jitsi version is 1.0.504-1 (I
know, pretty old) from stock repos on Debian Wheezy.

Best,
Stan