Hi Anuj, > I would like to know how to create an E-node and a D-node in > ML cluster environment?
Nobody seems to have answered your question yet. I'll give it a try. You may also be interested in this thread: http://markmail.org/thread/xtukevosgxllrskf, and perhaps also this one http://markmail.org/thread/63xxxqomamzxaxsv Any host can act as e-node, d-node or both. Whether it acts as a d-node is determined solely by the fact whether it is hosting forests or not. To become a e-node it only needs to be associated with a group that contains app servers (anyone, please correct me if I am wrong). It is not uncommon to have two groups in which one only contains e-nodes, and the other only contains d-nodes. Well, actually.. Databases aren't part of a group, but forest are hosted by hosts, and hosts have to be associated with a group. And it is legal to have one group with app servers that are associated with databases that are actually hosted by hosts in a different group, as long as the groups are in the same cluster, to make sure all hosts are aware of all groups and databases within the cluster.. > Should the port number for an XDBC server in one host/node be > same as that of other host/node of the same cluster? Each host is associated with one group, and one group only. App servers, like XDBC servers and HTTP servers, are configured within a group. All hosts that are associated with the same group, all host the same app servers. It requires a load-balancer though to make that transparent for the end user.. > How will we get the connection to a particular host in the > cluster for executing the x-query? Each host has it's own IP. You can use the IP combined with the port number of a particular app server to address one e-node specifically, but I recommend using a load-balancer to connect to any of the available e-nodes automatically. Connections to the necessary d-nodes are arranged on the background fully automatically. > How an x-query executes against these nodes? A request is typically handled on a single e-node, accessing as many d-nodes to gather the data it needs. > How to tell the MarkLogic, in the clustered environment, to > execute an x-query against a particular host/node in the cluster? See above. > Please reply. I am very eager to know how clustering works in > MarkLogic. I hope this helps. You might also want to read more about forest-level failover. MarkLogic does not depend on data-synchronisation to make scaling up far more easy, but uses a different strategy called forest-level failover. Good to be aware of, and interesting as well.. Kind regards, Geert Drs. G.P.H. Josten Consultant http://www.daidalos.nl/ Daidalos BV Source of Innovation Hoekeindsehof 1-4 2665 JZ Bleiswijk Tel.: +31 (0) 10 850 1200 Fax: +31 (0) 10 850 1199 http://www.daidalos.nl/ KvK 27164984 De informatie - verzonden in of met dit emailbericht - is afkomstig van Daidalos BV en is uitsluitend bestemd voor de geadresseerde. Indien u dit bericht onbedoeld hebt ontvangen, verzoeken wij u het te verwijderen. Aan dit bericht kunnen geen rechten worden ontleend. _______________________________________________ General mailing list [email protected] http://xqzone.com/mailman/listinfo/general
