[
https://issues.apache.org/jira/browse/IGNITE-14224?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17302265#comment-17302265
]
Vladimir Goncharov edited comment on IGNITE-14224 at 3/16/21, 11:25 AM:
------------------------------------------------------------------------
W10 HOME/PRO
JDK1.8.251/241 Oracle/Open
W/o fix - exception on spi.openSocket in ServerImpl.sendMessageDirectly.
IgniteSpiException : class org.apache.ignite.spi.IgniteSpiException: Failed to
send message to address [addr=/127.0.0.1:47500,
msg=TcpDiscoveryJoinRequestMessage [node=TcpDiscoveryNode
[id=b712cdf2-5e80-4d05-b45c-a113bdeea34b, consistentId=127.0.0.1:47501,
addrs=ArrayList [127.0.0.1], sockAddrs=HashSet [/127.0.0.1:47501],
discPort=47501, order=0, intOrder=0, lastExchangeTime=1615819551929, loc=true,
ver=2.11.0#20210310-sha1:e90f9666, isClient=false],
dataPacket=org.apache.ignite.spi.discovery.tcp.internal.DiscoveryDataPacket@33a630fa,
super=TcpDiscoveryAbstractMessage [sndNodeId=null,
id=3ec89563871-b712cdf2-5e80-4d05-b45c-a113bdeea34b, verifierNodeId=null,
topVer=0, pendingIdx=0, failedNodes=null, isClient=false]]]
Connection count doesnt increment and spi.joinTimeout is zero. This leads to
infintite cycle of cluster joining.
With fix - timeout exception after open socket and sending message as expected
witch correct stopping of joining node after 3 failed connection.
Got exception while starting (will rollback startup routine).
Command protocol successfully stopped: TCP binary
JDK 1.8.261/271/281 Oracle/Open
Works with and without patch as well
was (Author: shadowru):
Java 1.8.251/241 Oracle/Open
W/o fix - exception on spi.openSocket in ServerImpl.sendMessageDirectly.
IgniteSpiException : class org.apache.ignite.spi.IgniteSpiException: Failed to
send message to address [addr=/127.0.0.1:47500,
msg=TcpDiscoveryJoinRequestMessage [node=TcpDiscoveryNode
[id=b712cdf2-5e80-4d05-b45c-a113bdeea34b, consistentId=127.0.0.1:47501,
addrs=ArrayList [127.0.0.1], sockAddrs=HashSet [/127.0.0.1:47501],
discPort=47501, order=0, intOrder=0, lastExchangeTime=1615819551929, loc=true,
ver=2.11.0#20210310-sha1:e90f9666, isClient=false],
dataPacket=org.apache.ignite.spi.discovery.tcp.internal.DiscoveryDataPacket@33a630fa,
super=TcpDiscoveryAbstractMessage [sndNodeId=null,
id=3ec89563871-b712cdf2-5e80-4d05-b45c-a113bdeea34b, verifierNodeId=null,
topVer=0, pendingIdx=0, failedNodes=null, isClient=false]]]
Connection count doesnt increment and spi.joinTimeout is zero. This leads to
infintite cycle of cluster joining.
With fix - timeout exception after open socket and sending message as expected
witch correct stopping of joining node after 3 failed connection.
Got exception while starting (will rollback startup routine).
Command protocol successfully stopped: TCP binary
Java 1.8.261/271/281 Oracle/Open
Works with and without patch as well
> Extend logging of client failure handling.
> ------------------------------------------
>
> Key: IGNITE-14224
> URL: https://issues.apache.org/jira/browse/IGNITE-14224
> Project: Ignite
> Issue Type: Improvement
> Components: clients
> Affects Versions: 2.9.1
> Reporter: Stanilovsky Evgeny
> Assignee: Stanilovsky Evgeny
> Priority: Major
> Fix For: 2.11
>
> Time Spent: 1h 10m
> Remaining Estimate: 0h
>
> # "Finished serving remote node connection " but this message doesn't contain
> node id, it should be added (check
> org.apache.ignite.spi.discovery.tcp.ServerImpl.SocketReader#body).
> # When client worker is removed from clientMsgWorkers,
> org.apache.ignite.internal.util.IgniteUtils#closeQuiet(java.lang.AutoCloseable)
> is used to close the socket, we don't see any exception if it has happened,
> closeQuiet should be replaced with
> org.apache.ignite.internal.util.IgniteUtils#close(java.lang.AutoCloseable,
> org.apache.ignite.IgniteLogger) which works in the same way but prints
> exception.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
