Each node could be primary for some partitions and backup for other. It's defined by affinity function: https://apacheignite.readme. io/docs/affinity-collocation#section-affinity-function
Node, which performs fetch request can easily understand, to which partition key belongs and determine on which node it placed by using affinity function. Each node needs to store metadata since it's needed to perform operations on the cluster. Evgenii 2018-04-22 10:15 GMT+03:00 begineer <[email protected]>: > Hi, > I am bit confused about how ignite manages which nodes is primary and > secondary. Every distributed grid has some node which has information about > primary and secondary nodes so if primary goes down, one of primary is made > secondary which out any data loss(loss here means if primary node data > wan't > completely replicated in secondary before primary went down). How does it > work in ignite. > > 2. Also, If I send a request to fetch one entry of cache o a grid of 10 > nodes say. How does it work? Will the request land in random node, search > key there if not send it to next node and so on. Then if entry is found on > nth node, will the result be sent back to client from nth node or it comes > back to 1st node > > 3. Who stores the metadata information, as I understand, there is not > master > node in ignite. > > Thanks > > > > -- > Sent from: http://apache-ignite-users.70518.x6.nabble.com/ >
