[zeromq-dev] 0MQ multicast message lost?
ntupel at googlemail.com
ntupel at googlemail.com
Fri Oct 8 14:40:23 CEST 2010
On Fri, Oct 8, 2010 at 10:19 AM, Steven McCoy <steven.mccoy at miru.hk> wrote:
> Apply the patch from this email,
>
> [zeromq-dev] [PATCH] [Issue 96] Default 0MQ's PGM configuration to more
> lenient profile
> http://lists.zeromq.org/pipermail/zeromq-dev/2010-October/006881.html
Thanks,
I should have written, that I was still on zeromq-2.0.9 but now I
switched to git master and applied your patch. I also enabled trace
output as you advised previously. Unfortunately an assertion failure
occurs on receiver side (but it also occurs without your patch):
Minor: OpenPGM 5.0.78 () 2010-10-08 08:55:35 Linux x86_64
Minor: Using TSC timer.
Trace: Opening UDP encapsulated sockets.
Trace: Set socket sharing.
Trace: Request socket packet-info.
Trace: Assuming IP header size of 20 bytes
Trace: Assuming UDP header size of 8 bytes
Trace: Binding receive socket to INADDR_ANY
Trace: Binding send socket to interface index 2
Trace: Join multicast group 231.192.0.1 on interface index 2
Trace: Multicast send interface set to 10.0.11.109 index 2
Trace: Recv again on not-full
Trace: Recv again on not-full
Trace: Recv again on not-full
Trace: New pending data.
Trace: Recv again on not-full
Trace: New pending data.
...
Trace: Recv again on not-full
Trace: nak_rpt_expiry in 0.200000 seconds.
Trace: nak_rpt_expiry in 0.200000 seconds.
Trace: nak_rpt_expiry in 0.200000 seconds.
Trace: nak_rpt_expiry in 0.200000 seconds.
...
Trace: nak_rpt_expiry in 0.200000 seconds.
Trace: nak_rpt_expiry in 0.200000 seconds.
Trace: NAK backoff queue empty.
Trace: New pending data.
Trace: Recv again on not-full
Trace: New pending data.
Trace: Recv again on not-full
...
Trace: New pending data.
Trace: Recv again on not-full
Trace: nak_rpt_expiry in 0.200000 seconds.
Trace: nak_rpt_expiry in 0.200000 seconds.
Trace: nak_rpt_expiry in 0.200000 seconds.
Trace: NAK backoff queue empty.
Trace: NCF retry #322 attempt 1/50.
Trace: NCF retry #323 attempt 1/50.
...
Trace: NCF retry #359 attempt 1/50.
Trace: NCF retry #360 attempt 1/50.
Trace: NCF retry #472 is delayed 0.200000 seconds.
Trace: Next expiry set in 0.200000 seconds.
Trace: New pending data.
Trace: Recv again on not-full
...
Trace: New pending data.
Trace: Recv again on not-full
Assertion failed: false (poller_base.cpp:64)
Aborted
For completeness this is logged on sender side:
Minor: OpenPGM 5.0.78 (exported) 2010-10-08 10:46:33 Linux x86_64
Minor: Using TSC timer.
Trace: Opening UDP encapsulated sockets.
Trace: Set socket sharing.
Trace: Request socket packet-info.
Trace: Assuming IP header size of 20 bytes
Trace: Assuming UDP header size of 8 bytes
Trace: Create transmit window.
Trace: Binding receive socket to INADDR_ANY
Trace: Binding send socket to interface index 3
Trace: Setting rate regulation to 512000 bytes per second.
Trace: Join multicast group 231.192.0.1 on interface index 3
Trace: Multicast send interface set to 10.0.11.110 index 3
Trace: Recv again on not-full
Trace: Recv again on not-full
Trace: Recv again on not-full
Trace: Recv again on not-full
Trace: Recv again on not-full
Trace: Failed to push retransmit request for #322
Trace: Recv again on not-full
Trace: Recv again on not-full
Trace: Failed to push retransmit request for #472
Trace: Recv again on not-full
Trace: Closing receive socket.
Trace: Closing send socket.
Trace: Flushing PGM source with session finish option broadcast SPMs.
Trace: Destroying transmit window.
Trace: Destroying rate control.
Trace: Closing send with router alert socket.
This is with remote_thr and local_thr where rate is set to 4096.
Thanks,
nt
More information about the zeromq-dev
mailing list