[jboss-jira] [JBoss JIRA] (JGRP-2485) UDP is not working after upgarde to 3_6_19 from jgroups-3.4.0.Alpha2

[Rashmi Acharya (Jira)]

    [ https://issues.redhat.com/browse/JGRP-2485?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14187203#comment-14187203 ] 

Rashmi Acharya commented on JGRP-2485:
--------------------------------------

Hi Bela

We are not able to upgrade to supported version of Jgroups with this error. We have many customers who are using UDP for their cluster communication and without understanding the root cause and underlying architectute change we are not able to make progress on code side.

 

Your help is really required on why we are not able to make it work with old configurations and how can we debug this issue further.

 

Thanks in advance !

> UDP is not working after upgarde to 3_6_19 from jgroups-3.4.0.Alpha2
> --------------------------------------------------------------------
>
>                 Key: JGRP-2485
>                 URL: https://issues.redhat.com/browse/JGRP-2485
>             Project: JGroups
>          Issue Type: Bug
>    Affects Versions: 3.6.19
>            Reporter: Janardhan Naidu
>            Assignee: Bela Ban
>            Priority: Major
>         Attachments: noapp.log.D20200619.T053202_NODE1
>
>
> Hi Team,
>  
> we just upgraded from  jgroups-3.4.0.Alpha2 to 3_6_19. post the UDP cluster communication is not working.
> after upgrade, we were hitting some warning as below and we solved those by changing property_string of UDP.
> *Warnings:*
> *WARNING: JGRP000014: Discovery.timeout has been deprecated: GMS.join_timeout should be used instead*
>  [2020-06-17 14:05:49.271] ALL 000000000000 GLOBAL_SCOPE Jun 17, 2020 2:05:49 PM org.jgroups.stack.Configurator resolveAndAssignField
>  *WARNING: JGRP000014: Discovery.num_initial_members has been deprecated: will be ignored*
> [2020-06-17 14:05:49.396] ALL 000000000000 GLOBAL_SCOPE Jun 17, 2020 2:05:49 PM org.jgroups.protocols.pbcast.NAKACK init
>  *WARNING: use_mcast_xmit should not be used because the transport (TCP) does not support IP multicasting; setting use_mcast_xmit to false*
>  
> To solve above warnings, we went through tcp.xml which was shipped using in 3_6_19 jar:
> and now our properties looks like:
> *property_string=UDP*(bind_addr=*hostIP*;bind_port=39061;mcast_addr=239.255.166.17;mcast_port=39060;ip_ttl=32;mcast_send_buf_size=150000;mcast_recv_buf_size=80000):PING:MERGE2(min_interval=5000;max_interval=10000):FD_ALL(interval=5000;timeout=20000):VERIFY_SUSPECT(timeout=1500):pbcast.NAKACK(use_mcast_xmit=true;retransmit_timeout=300,600,1200,2400,4800;discard_delivered_msgs=true):UNICAST:pbcast.STABLE(desired_avg_gossip=20000):FRAG(frag_size=8096):pbcast.GMS(join_timeout=5000;print_local_addr=true)
>  
> *distribution_property_string=TCP*(bind_port=39060;thread_pool_rejection_policy=run):TCPPING(async_discovery=true;initial_hosts=*hostIP*[39060];port_range=0;):MERGE2(min_interval=3000;max_interval=5000):FD_SOCK:FD(timeout=5000;max_tries=48):VERIFY_SUSPECT(timeout=1500):pbcast.NAKACK(use_mcast_xmit=false;discard_delivered_msgs=true;):pbcast.STABLE(stability_delay=1000;desired_avg_gossip=20000;max_bytes=0):pbcast.GMS(join_timeout=5000;print_local_addr=true)
>  
> *lock.protocolStack=UDP*(bind_addr=*hostIP*;bind_port=39062;mcast_addr=239.255.166.17;mcast_port=39069;ip_ttl=32;mcast_send_buf_size=150000;mcast_recv_buf_size=80000):PING:MERGE2(min_interval=5000;max_interval=10000):FD_ALL(interval=5000;timeout=20000):VERIFY_SUSPECT(timeout=1500):pbcast.NAKACK(use_mcast_xmit=true;retransmit_timeout=300,600,1200,2400,4800):UNICAST:pbcast.STABLE(desired_avg_gossip=20000):FRAG(frag_size=8096):pbcast.GMS(join_timeout=5000;print_local_addr=true):CENTRAL_LOCK(num_backups=2)
>  
>  
> With above properties, we are not seeing any warning or error. But still cluster communication is not working with UDP. 
> Please help me in resolving the same 
>  
> Thank 
> Janardhan



--
This message was sent by Atlassian Jira
(v7.13.8#713008)