[jboss-jira] [JBoss JIRA] (JGRP-2335) JDBC_PING: code for determining the coordinator hangs in certain conditions

[Bela Ban (Jira)]

    [ https://issues.jboss.org/browse/JGRP-2335?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13766913#comment-13766913 ] 

Bela Ban commented on JGRP-2335:
--------------------------------

Re your suggestions:
a) somehow involve MembershipChangePolicy in ordering strategy

This is something you can do in your application. Both view creation _and_ address creation are pluggable and can be done is user code.

b) make the new server (joiner) the first in the sorted set, or

See a)

c) make UUID addresses to sort depending on their time of creation.

This can also be done with custom addresses (see above). Requirement of course is the same wall clock time everywhere, which requires time synchronization.

I'm closing this issue for now. Please feel free to reopen with exact steps to reproduce.

> JDBC_PING: code for determining the coordinator hangs in certain conditions
> ---------------------------------------------------------------------------
>
>                 Key: JGRP-2335
>                 URL: https://issues.jboss.org/browse/JGRP-2335
>             Project: JGroups
>          Issue Type: Bug
>            Reporter: Aieksiei Illarionov
>            Assignee: Bela Ban
>            Priority: Major
>             Fix For: 4.1.3
>
>
> Affected version:
> {code:xml}
>         <dependency>
>             <groupId>org.jgroups</groupId>
>             <artifactId>jgroups</artifactId>
>             <version>4.0.0.Final</version>
>         </dependency>
> {code}
> ClientGmsImpl#joinInternal hangs because #firstOfAllClients always returns false when all of the following conditions are satisfied:
> - using JDBC_PING for discovery protocol
> - JGROUPSPING table contains data from previous sessions
> - all of the previous sessions were killed (kill -9)
> - AddressGenerator is not customized
> The sorted set
> {code:java}
> SortedSet<Address> clients=new TreeSet<>();
> {code}
> contains the dead servers discovered from JGROUPSPING. When the new server is added to the sorted set, it never becomes the first in the sorted set.
> Suggestions: either
> a) somehow involve MembershipChangePolicy in ordering strategy, or
> b) make the new server (joiner) the first in the sorted set, or
> c) make UUID addresses to sort depending on their time of creation.
> I've used the following config:
> {code:xml}
> <!--
>     TCP based stack, with flow control and message bundling. This is usually used when IP
>     multicasting cannot be used in a network, e.g. because it is disabled (routers discard multicast).
>     Note that TCP.bind_addr and TCPPING.initial_hosts should be set, possibly via system properties, e.g.
>     -Djgroups.bind_addr=192.168.5.2 and -Djgroups.tcpping.initial_hosts=192.168.5.2[7800]
>     author: Bela Ban
> -->
> <config xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
>         xmlns="urn:org:jgroups"
>         xmlns:fork="fork"
>         xsi:schemaLocation="urn:org:jgroups http://www.jgroups.org/schema/jgroups.xsd fork http://www.jgroups.org/schema/fork-stacks.xsd">
>     <TCP bind_port="7800"
>          port_range="10"
>          bind_addr="<placeholder here>"
>          recv_buf_size="${tcp.recv_buf_size:130k}"
>          send_buf_size="${tcp.send_buf_size:130k}"
>          max_bundle_size="64K"
>          sock_conn_timeout="300"
>          thread_pool.min_threads="0"
>          thread_pool.max_threads="20"
>          thread_pool.keep_alive_time="30000"/>
>     <JDBC_PING
>             remove_all_data_on_view_change="true"
>             connection_driver="com.microsoft.sqlserver.jdbc.SQLServerDriver"
>             connection_url="jdbc:sqlserver://localhost:1433;databaseName=mydatabase"
>             connection_username="user"
>             connection_password="password"
>     />
>     <MERGE3  min_interval="10000"
>              max_interval="30000"/>
>     <FD_SOCK/>
>     <FD timeout="3000" max_tries="3" />
>     <VERIFY_SUSPECT timeout="1500"  />
>     <BARRIER />
>     <pbcast.NAKACK2 use_mcast_xmit="false"
>                     discard_delivered_msgs="true"/>
>     <UNICAST3
>             conn_close_timeout="240000"
>             xmit_interval="5000"/>
>     <pbcast.STABLE desired_avg_gossip="50000"
>                    max_bytes="4M"/>
>     <pbcast.GMS print_local_addr="true" join_timeout="2000"
>                 view_bundling="true"
>                 membership_change_policy="ru.illar.AppMembershipChangePolicy"/>
>     <MFC max_credits="2M"
>          min_threshold="0.4"/>
>     <FRAG2 frag_size="60K"  />
>     <!--RSVP resend_interval="2000" timeout="10000"/-->
>     <pbcast.STATE_TRANSFER/>
> </config>
> {code}



--
This message was sent by Atlassian Jira
(v7.12.1#712002)