[javagroups-users] Fw: Strange behaviour

[Dima G. <di...@ma...>]

Anyone ?

----- Original Message ----- 
From: "Dima Gutzeit" <di...@ma...>
To: <jav...@li...>
Sent: Thursday, March 17, 2005 19:48
Subject: Strange behaviour


> Hi,
>
> I've been using the latest version from JGroups repository and I have one
> strange thing.
> I've tested my application on four different computer, and only on one 
> I've
> experiencing the following :
>
> After connecting to a channel I see the following :
>
> In the jgroups logs :
>
> [GMS]  WARN 2005-03-17 19:36:24,016 checkSelfInclusion() failed,
> 192.168.0.70:45779 is not a member of view [192.168.0.200:33836|11422]
> [192.168.0.200:33836]; discarding view
> [JChannel]  INFO 2005-03-17 19:36:31,234 received an EXIT event, will 
> leave
> the channel
> [JChannel]  INFO 2005-03-17 19:36:31,235 closing the channel
> [JChannel]  INFO 2005-03-17 19:36:32,053 reconnecting to group
> TCIManagementGroup
> [UDP]  INFO 2005-03-17 19:36:32,071 sockets will use interface 
> 192.168.0.70
> [UDP]  INFO 2005-03-17 19:36:32,101 socket information:
> local_addr=192.168.0.70:45780, mcast_addr=228.8.8.8:45566,
> bind_addr=/192.168.0.70, ttl=32
> sock: bound to 192.168.0.70:45780, receive buffer size=64000, send buffer
> size=32000
> mcast_recv_sock: bound to 192.168.0.70:45566, send buffer size=131071,
> receive buffer size=80000
> mcast_send_sock: bound to 192.168.0.70:45781, send buffer size=131071,
> receive buffer size=80000
>
> It keeps hapenning every 10 seconds.
>
> And in the system output I see :
>
> -------------------------------------------------------
> GMS: address is 192.168.0.70:45779
> -------------------------------------------------------
>
> -------------------------------------------------------
> GMS: address is 192.168.0.70:45780
> -------------------------------------------------------
>
> -------------------------------------------------------
> GMS: address is 192.168.0.70:45782
> -------------------------------------------------------
>
> -------------------------------------------------------
> GMS: address is 192.168.0.70:45784
> -------------------------------------------------------
>
> -------------------------------------------------------
> GMS: address is 192.168.0.70:45786
> -------------------------------------------------------
>
> -------------------------------------------------------
> GMS: address is 192.168.0.70:45788
> -------------------------------------------------------
>
> -------------------------------------------------------
> GMS: address is 192.168.0.70:45790
> -------------------------------------------------------
>
> -------------------------------------------------------
> GMS: address is 192.168.0.70:45792
> -------------------------------------------------------
>
> -------------------------------------------------------
> GMS: address is 192.168.0.70:45794
> -------------------------------------------------------
>
> -------------------------------------------------------
> GMS: address is 192.168.0.70:45796
> -------------------------------------------------------
>
>
>
> On other computers I've tested exactly the same application and it 
> connects
> only once to a channel and stays connected.
> When I have another computer connected on that channel I see that new 
> member
> joins (the problematic computer) and immodestly departs ?!
>
> I am using the following protocol stack:
>
> UDP(bind_addr={0};mcast_addr={1};mcast_port={2};ip_ttl=32;\
>
> mcast_send_buf_size=150000;mcast_recv_buf_size=80000):\
>
> PING(timeout=1250;num_initial_members=2):\
>
> MERGE2(max_interval=10000;min_interval=5000):\
>                                        FD_SOCK:\
>                                        FD(timeout=1000;max_tries=2):\
>                                        VERIFY_SUSPECT(timeout=500):\
>
> pbcast.NAKACK(gc_lag=50;retransmit_timeout=300):\
>                                        UNICAST(timeout=1000):\
>
> pbcast.STABLE(desired_avg_gossip=5000):\
>
> FRAG(frag_size=4096;down_thread=false;up_thread=false):\
>
> pbcast.GMS(join_timeout=5000;join_retry_timeout=2000;\
>                                        shun=false;print_local_addr=true):\
>                                        pbcast.STATE_TRANSFER
>
>
> Does anyone has a clue on why it may happen ?
>
> Thanks in advance.
>
> Dima Gutzeit.
>