CODEWITHSUNDEEP
erlangmnesiadistributed-system
jasmad
jasmad

Reputation: 115

Mnesia - Replicate ram_copy table to disc_only_copy table from another node

What I want to do is this:

Create a table ram_copies in node A which resides in a computer with lot of Ram and CPU power

mnesia:create_table(person, [{ram_copies, [node()]},
                        {frag_properties, [{node_pool, [node()]}, {n_fragments, 10}, {n_ram_copies, 1}]},
                        {attributes, record_info(fields, person)},
                        {storage_properties, [{ets, [compressed, {write_concurrency,true}, {read_concurrency,true}]}]}])

And node A starts working this ram_copy table.

At some unspecified time (which may be 1 second or 1 hour) after the node "A" have created the table, create node B on a computer with lots of hard drive, but less RAM and CPU power , to replicate (structure and data) RAM table in the "a" node but disc_only_copies.

The idea is to run critical processes using all the power at node A, and run trivial processes on another node and keep the data synchronized.

Is this possible?

Greetings.

Upvotes: 2

Views: 861

Answers (1)

Hamidreza Soleimani
Hamidreza Soleimani

Reputation: 2544

Yes, it is possible.

Let's say we have two nodes; node_1 which has lots of Disc and node_2 which has lots of RAM, and we have a foo table which must be disc_only_copies in node_1 and ram_copies in node_2. The steps are as follows:

Node 1:

$ erl -sname node_1
(node_1@host)1> mnesia:create_schema([node()]). 
(node_1@host)2> mnesia:start(). 
(node_1@host)3> mnesia:add_table_copy(schema, node_2@host, ram_copies).

Node 2:

$ erl -sname node_2
(node_2@host)1> mnesia:start().
(node_2@host)2> mnesia:change_config(extra_db_nodes, [node_1@host]).

That's it. Now we can create foo table in node_1:

Node 1:

(node_1@host)4> mnesia:create_table(foo, [
                    {disc_only_copies, [node_1@host]}, 
                    {ram_copies, [node_2@host]}]).

Finally we can check if it is created on both nodes with mnesia:info/0:

Node 1:

running db nodes   = [node_2@host,node_1@host]
stopped db nodes   = [] 
master node tables = []
remote             = []
ram_copies         = []
disc_copies        = [schema]
disc_only_copies   = [foo]
[{node_1@host,disc_copies},{node_2@host,ram_copies}] = [schema]
[{node_1@host,disc_only_copies},{node_2@host,ram_copies}] = [foo]

Node 2:

running db nodes   = [node_1@host,node_2@host]
stopped db nodes   = [] 
master node tables = []
remote             = []
ram_copies         = [foo,schema]
disc_copies        = []
disc_only_copies   = []
[{node_1@host,disc_copies},{node_2@host,ram_copies}] = [schema]
[{node_1@host,disc_only_copies},{node_2@host,ram_copies}] = [foo]

Upvotes: 3

Related Questions