BTW: this is the result of netstat on the zk server:
tcp 0 0 0.0.0.0:8080 0.0.0.0:* LISTEN 1838/ java tcp 0 0 ***.***.***.***:3888 0.0.0.0:* LISTEN 1214/ java tcp 0 0 0.0.0.0:40864 0.0.0.0:* LISTEN 1214/ java tcp 0 0 0.0.0.0:2181 0.0.0.0:* LISTEN 1214/ java as you can see, the zk server is listening on port 3888 as specified in the zk config file and on 2181 as specified in the RSA discovery settings in karaf while this is the result of netstat on the karaf server that is using zk to communicate with remote services. tcp 0 0 127.0.0.1:1099 0.0.0.0:* LISTEN 15258/ java tcp 0 0 0.0.0.0:46445 0.0.0.0:* LISTEN 15258/ java tcp 0 0 127.0.0.1:38707 0.0.0.0:* LISTEN 15258/ java tcp 0 0 0.0.0.0:8181 0.0.0.0:* LISTEN 15258/ java tcp 0 0 0.0.0.0:2550 0.0.0.0:* LISTEN 15258/ java tcp 0 0 127.0.0.1:44444 0.0.0.0:* LISTEN 15258/ java tcp 0 0 127.0.0.1:32000 0.0.0.0:* LISTEN 15258/ java tcp 0 0 0.0.0.0:8101 0.0.0.0:* LISTEN 15258/ java --- Schéin Gréiss, Mit freundlichen Grüßen, Meilleures salutations, Kind regards, Alex Weirig Responsable Technique Ville de Luxembourg - Centre Technolink 2, rue Charles de Tornaco L - 2623 LUXEMBOURG [ mailto:alex.wei...@technolink.lu | alex.wei...@technolink.lu ] Tel: +352 4796 - 6127 Fax: +352 42 88 81 [ http://www.technolink.lu/ | www.technolink.lu ] From: "Alex Weirig" <alex.wei...@technolink.lu> To: "user" <user@aries.apache.org> Sent: Friday, September 15, 2017 11:55:23 PM Subject: Re: Echo TCP example not working on computer with multiple NICs Jonathan, good to hear that it is somewhat working although not the way you expected ... Are you sure you want to run zk on the same servers as your servicemix? If you get other services relying on zk you might want to run zk on dedicated "servers" (that can be VMs or nowadays containers). Since zk doesn't require a lot of resources it's not that bad to have multiple isolated copies of zk running. As I said we went with 3 instances (one of them always being elected to be the master). The zk configuration file is residing on a shared drive mounted into each of the zk instances, so they all share the same configuration. We monitor the zk with exhibitor. I wonder if you didn't see the server on the consumer computer because that zk doesn't have anything to publish while the service computer is actually publishing a service. I can try and check what the netstat looks like on our servers. I remember my early zk days when I ran zk on the same server as karaf and also had the bundle stop and start the zk service, so I think that's expected. At the end of the day, you only need to run zk on one single server. Your "Service" computer needs to know where zk is running to publish it's services to zk, your client computer needs to know where zk is running to create a "local" proxy for the remote service when it is used locally on the client computer. Basically these are the bundles I'm running in Karaf to use zk with fastbin transport. 106 │ Active │ 80 │ 1.10.0 │ Aries Remote Service Admin Core 107 │ Active │ 80 │ 1.10.0 │ Aries Remote Service Admin Discovery Gogo Commands 108 │ Active │ 80 │ 1.10.0 │ Aries Remote Service Admin Discovery Local 109 │ Active │ 80 │ 1.10.0 │ Aries Remote Service Admin Discovery Zookeeper 110 │ Active │ 80 │ 1.10.0 │ Aries Remote Service Admin Event Publisher 111 │ Active │ 80 │ 1.10.0 │ Aries Remote Service Admin provider FastBin 112 │ Active │ 80 │ 1.10.0 │ Aries Remote Service Admin SPI 113 │ Active │ 80 │ 1.10.0 │ Aries Remote Service Admin Topology Manager 127 │ Active │ 80 │ 3.4.7 │ ZooKeeper Bundle Gogo commands are not mandatory I think but can be useful. You can also deploy a different provider other than fastbin. --- Schéin Gréiss, Mit freundlichen Grüßen, Meilleures salutations, Kind regards, Alex Weirig Responsable Technique Ville de Luxembourg - Centre Technolink 2, rue Charles de Tornaco L - 2623 LUXEMBOURG [ mailto:alex.wei...@technolink.lu | alex.wei...@technolink.lu ] Tel: +352 4796 - 6127 Fax: +352 42 88 81 [ http://www.technolink.lu/ | www.technolink.lu ] From: "Jonathan knez" <jonathank...@hotmail.com> To: "Alex Weirig" <alex.wei...@technolink.lu>, user@aries.apache.org Sent: Friday, September 15, 2017 10:56:55 PM Subject: Re: Echo TCP example not working on computer with multiple NICs Alex, Thank you for the help. This got me pointed in the right direction. Although there are still some things I don't understand, it is working now. I downloaded the zookeeper release archive from their website. I then ran the zkCli script on each of my servers and attempted to connect on the default port of 2181. This was successful on the "Service" computer. It failed on the "Consumer" computer. I then ran "sudo netstat -anp | grep LISTEN | grep java" to see what ports java was even listing on. Port 2181 is not one of them. So I can say with confidence that the zookeeper server is not running on the "Consumer" computer. I ran the netstat on the "Service" computer and I see 2181 in the list. When I stop the "Aries Remote Service Admin Discovery Zookeeper" bundle in servicemix, the zookeeper server stops. When I start that bundle, the zookeeper server starts. So it looks like that bundle is actually launching the zookeeper server. I verified that bundle is started on the "Consumer" computer and it is. I can start and stop that bundle all I want but zookeeper never successfully starts. This is the part I don't understand. Regardless, I did what you said and filled out <service_mix_dir>/etc/org.apache.aries.rsa.discovery.zookeeper.cfg with the host and port for my "Service" computer. So both computers are sharing the same one zookeeper instance. With that change, the echo example works correctly. So the problem was down to the fact that zookeeper was not running on one computer. I still don't know why it was not running, but I was able to work around it. I suppose in the long run I will want to run one instance of zookeeper per computer, regardless of how many servicemix instances I am running on that computer. With that in mind, I probably wont' ever take the time to figure out why zookeeper was not running in servicemix on that computer. Jonathan From: Alex Weirig <alex.wei...@technolink.lu> Sent: Friday, September 15, 2017 3:44 AM To: user@aries.apache.org; Jonathan knez Subject: Re: Echo TCP example not working on computer with multiple NICs Hi Jonathan, I'm not sure if this can help you but this is how I use karaf and zookeeper: 1) tell RSA where to find zookeeper: Place a file named "org.apache.aries.rsa.discovery.zookeeper.cfg" with the following content into karaf's "etc" folder: zookeeper.host = dns name of your zookeeer server zookeeper.port = 2181 zookeeper.timeout = 3000 2) when using fastbin as transport I have to create a file named "org.apache.aries.rsa.provider.fastbin.cfg" in the "etc" folder containing: uri = tcp://0.0.0.0:2550 otherwise I get a port collision 3) configure zookeeper, e.g. # The number of milliseconds of each tick tickTime=2000 # The number of ticks that the initial # synchronization phase can take initLimit=10 # The number of ticks that can pass between # sending a request and getting an acknowledgement syncLimit=5 # the directory where the snapshot is stored. # do not use /tmp for storage, /tmp here is just # example sakes. dataDir=/opt/zookeeper-data # Transaction log #dataLogDir=/opt/zookeeper-data-log # the port at which the clients will connect clientPort=2181 # the maximum number of client connections. # increase this if you need to handle more clients maxClientCnxns=60 # # Be sure to read the maintenance section of the # administrator guide before turning on autopurge. # # [ http://zookeeper.apache.org/doc/current/zookeeperAdmin.html#sc_maintenance | http://zookeeper.apache.org/doc/current/zookeeperAdmin.html#sc_maintenance ] # # The number of snapshots to retain in dataDir #autopurge.snapRetainCount=3 # Purge task interval in hours # Set to "0" to disable auto purge feature #autopurge.purgeInterval=1 # Cluster servers server.1=zk1:2888:3888 server.2=zk2:2888:3888 server.3=zk3:2888:3888 Schéin Gréiss, Mit freundlichen Grüßen, Meilleures salutations, Kind regards, Alex Weirig Responsable Technique Ville de Luxembourg - Centre Technolink 2, rue Charles de Tornaco L - 2623 LUXEMBOURG [ mailto:alex.wei...@technolink.lu | alex.wei...@technolink.lu ] Tel: +352 4796 - 6127 Fax: +352 42 88 81 [ http://www.technolink.lu/ | www.technolink.lu ] On 15/09/2017 00:49, Jonathan knez wrote: I am just getting started with remote services in OSGi. To that end, I have downloaded Apache Karaf 4.1.2 and run through the Echo TCP example from the Aries-RSA github page. Here is a link. [ https://github.com/apache/aries-rsa/tree/master/examples/echotcp | https://github.com/apache/aries-rsa/tree/master/examples/echotcp ] The example ran successfully when running two karaf instances on my single computer. I then ran one karaf instance on another networked computer. The two computers are connected via a LAN that allows all manner of network communication including multicast. When running the "Service" on one computer and the "Consumer" on the other computer, the example does not work. The Consumer bundle stays in a Waiting state. When I run rsa:endpoints on the "Service" computer, I see that the EchoService is correctly bound to a local IP address. When I run rsa:endpoints on the "Consumer" computer, it shows no services. The issue appears to be that the karaf instance on the "Consumer" computer has not found the "Service" endpoint on the other computer. Both computers have multiple NICs and multiple active network connections. Only one connection allows them to talk to each other. It just so happens that the "Service" endpoint is using the right address. I suspect that the "Consumer" karaf instance is not so lucky and is using the wrong NIC. So my question is how can I configure this? I believe this might be a Zookeeper question but I am not clear on how to configure the zookeeper instance that Aries-RSA launches. I appreciate the help. Jonathan -- This email was Anti Virus checked by SOPHOS UTM
