Hello!
There is a known issue with Docker gateway - it is present on all machines,
leading Ignite nodes to believe they can reach each other via this
interface.
Perhaps the same reasoning for other addresses.
You can set IgniteConfiguration.localHost or
TcpCommunicationSpi.localAddress to the desired address to mitigate this.
The caveat is, you have to know specific local address to supply it in
configuration.
Regards,
--
Ilya Kasnacheev
сб, 9 нояб. 2019 г. в 03:11, Conrad Mukai (cmukai) :
> I am running a 4 node ignite cluster in AWS. We are running ignite inside
> of Docker containers (the image is apacheignite/ignite:2.7.6 from
> Dockerhub). The container is setup to use the host network.
>
>
>
> I noticed a lot of warning from the tcp-comm-worker-#1 thread in which it
> is attempting to connect to unrecognized IP addresses:
>
>
>
> [2019-11-08 22:38:24,205][WARN ][tcp-comm-worker-#1][TcpCommunicationSpi]
> Connect timed out (consider increasing 'failureDetectionTimeout'
> configuration property) [addr=ip-172-17-0-1.ec2.internal/172.17.0.1:47104,
> failureDetectionTimeout=30]
>
> [2019-11-08 22:40:31,501][WARN ][tcp-comm-worker-#1][TcpCommunicationSpi]
> Connect timed out (consider increasing 'failureDetectionTimeout'
> configuration property) [addr=/9.0.1.1:47104,
> failureDetectionTimeout=30]
>
> [2019-11-08 22:44:40,973][WARN ][tcp-comm-worker-#1][TcpCommunicationSpi]
> Connect timed out (consider increasing 'failureDetectionTimeout'
> configuration property) [addr=/198.51.100.3:47105,
> failureDetectionTimeout=30]
>
> [2019-11-08 22:45:03,931][WARN ][tcp-comm-worker-#1][TcpCommunicationSpi]
> Connect timed out (consider increasing 'failureDetectionTimeout'
> configuration property) [addr=ip-172-17-0-1.ec2.internal/172.17.0.1:47104,
> failureDetectionTimeout=30]
>
> [2019-11-08 22:47:11,245][WARN ][tcp-comm-worker-#1][TcpCommunicationSpi]
> Connect timed out (consider increasing 'failureDetectionTimeout'
> configuration property) [addr=ip-172-31-254-1.ec2.internal/
> 172.31.254.1:47104, failureDetectionTimeout=30]
>
> [2019-11-08 22:50:00,333][WARN ][tcp-comm-worker-#1][TcpCommunicationSpi]
> Connect timed out (consider increasing 'failureDetectionTimeout'
> configuration property) [addr=/198.51.100.1:47109,
> failureDetectionTimeout=30]
>
>
>
> These IP addresses were not configured in the discoverySpi. I recognize
> 172.17.0.1 as part of the default Docker bridge network, but as I mentioned
> earlier we are using the host network. Also 172.31.0.0/16 is the default
> VPC in AWS, but we are not using that VPC. I have no idea why these IP
> addresses are being accessed. Can someone explain what is going on?
>
>
>
> Thanks in advance,
>
> Conrad
>
>
>
