Start your new node (b@node
) erl -sname b -mnesia dir '"/path/to/storage"' -s mnesia
. This starts a new ram_copies node called b@node
.
On your original node (a@node
), at the erlang prompt execute mnesia:change_config(extra_db_nodes, ['b@node']).
This will cause the original node to connect b
to the mnesia cluster. At this point, b@node
has joined the cluster but only has a copy of the schema.
To make new the node b@node
capable of storing disc copies, we need to change the schema table type on b@node
from ram_copies
to disc_copies
. Run mnesia:change_table_copy_type(schema, 'b@node', disc_copies).
on any node.
b@node
only has a copy of the schema at this point. To copy all the tables from a@node
to b@node
and maintain table types, you can run:
[{Tb, mnesia:add_table_copy(Tb, node(), Type)}
|| {Tb, [{'a@node', Type}]} <- [{T, mnesia:table_info(T, where_to_commit)}
|| T <- mnesia:system_info(tables)]].
This command may take a while to execute as it will copy the contents of each table over the network.
b@node
is now an exact replica of a@node
. You could modify that statement - replace the Type
variable with disc_only_copies
in the call to mnesia:add_table_copy/3
in order to copy the tables but ensure they're on disc only.
The mnesia documentation explains how to use the functions I've shown here.
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…