i have damaged mnesia database beyond repair result of overestimating fragility of implementation. when try mnesia api records need not visible though keys visible in file. though documentation indicates mnesia artifacts dets files cannot opened or identified dets artifacts. ps: dump_to_textfile() not work either.
eventually able dump db. did not end mnesia problems gave me options did not have before.
setup: had implemented master-master mnesia cluster. (read docs). turns out not seasoned erlang programmer uses mnesia replication there many flaws. in fact come information erlang inner circle , few l1 teams too. in case, however, work in production. , that's when problems started.
we started getting db consistency errors and, favorite, network or db partition errors. takes highly skilled , knowledgeable individual recover lot of planning , code in advance; did not have.
ultimately took 2 steps. (a) removed second app though db in master-master cluster; 1 slave because never used master. (b) in second implementation split cluster app ran on single node single db. #a in production , #b warm standby. replication manual writes rare.
in single node deployment there 2 nodes. first node application; app@ks , on same hardware "erl" node when needed rpc app , see how things going.
my solution: when posted question trying dump contents of mnesia db. having number of problems because trying access db admin node application node operational.
because trying access mnesia lib erl node db not local erl node , dump_to_textfile produced empty file. had success when used rpc tell app@ks node dump.
still undefined when launched admin node set mnesia dir parameter same folder app@ks node. have vague memory undesirable.
there many more mnesia issues solve none refer problem reported. still not know how extract raw data various db files.
Comments
Post a Comment