Plasma GitLab Archive
Projects Blog Knowledge

Plasma Project:

Class type Nn_config.nn_node_config

class type nn_node_config = object .. end

method nn_clustername : string
The clustername
method nn_hostname : string
The hostname of this node. Must be resolvable to a reasonable IP address so that other nodes can reach this node.
method nn_nodes : string list
The name nodes in "host:port" syntax
method nn_node_timeout : float
The timeout for connects and RPC calls to name nodes
method nn_node_alive_min : int
How many name nodes need to be alive to continue with operations
method nn_node_alive_min_startup : int
How many name nodes need to be alive at cluster startup
method nn_rank : string
The rank of this node compared to other nodes. At startup time, the node with the lowest rank is chosen as coordinator. Each node must have a distinct rank. Ranks are compared as in
method nn_elect_timeout : float
For how many seconds to wait for neighbor namenodes
method nn_inodecaches : string list
Which inodecaches are announced to the clients
method nn_replication : int
Default replication factor
method nn_lock_timeout : float
After these many seconds transactions are not retried on ECONFLICT
method nn_pool_size : int
How much pool RAM is allocated (shared memory)
method nn_consistency_checks : bool
Whether to do more consistency checks, even the time-consuming ones
method dn_blocksize : int
Expected blocksize of datanodes
method dn_port : int
The data node port
method dn_node_timeout : float
The timeout for connects and RPC calls to data nodes
method dn_security_level : Pfs_auth.level
The security level the datanodes expect
method dn_discovery_addrs : Unix.inet_addr list
The address to use for discovering datanodes. Normally [], i.e. multicasting is enabled for this.
method dn_discovery_period : float
Every how many seconds a new discovery messge is emitted
method dn_multicast_ttl : int
The number of hops (routers+1) for multicasting. Default is 1
method dn_max_errors : int
For the liveliness monitoring: how many errors (missing response messages) are at most acceptable. If more errors occur, the datanode is considered as dead.
This web site is published by Informatikbüro Gerd Stolpmann
Powered by Caml