clean up migrations scripts
Opened this issue · 3 comments
0xmountaintop commented
Tejón Melero (@gcomte, tg hanle: subatomiq): The file 20210114140803_kline.sql exists in the 'migrations' folder as well as in the 'migrations/ts' folder. These files are identical. Why does it exist twice? I see the one in the 'ts' folder is being called in persist.rs, but does it even make sense to call this file multiple times? --> It contains CREATE TABLE queries ...
0xmountaintop commented
- make it clear that the purpose of db_log and db_history
- sort out what should be in "migrations/ts" and what should not be
0xmountaintop commented
- we use
db_log
for operation_log (and slices) and config (assets/markets...) - we use
db_history
for states (history) and kline
MIGRATORs are a bit messy:
- message::persist::MIGRATOR is related to persistor bin, and reads
"./migrations/ts"
- then seems we should also put all
db_history
stuff into"./migrations/ts"
?
- then seems we should also put all
- crate::persist may write to MQ or write to DB directly (so it connects both
db_log
anddb_history
), and reads"./migrations"
0xmountaintop commented
we should either fully combine these 2 DBs, or
use different endpoint for these 2 DBs and let them do their own jobs.
so that
db_log
is involved with "operation_log (and slices) and config (assets/markets...)"-related scripts in "./migrations/
"db_history
is involved with- "state_changes"-related scripts in "
./migrations/
" and - "kline"-related scripts in "
./migrations/ts
"
- "state_changes"-related scripts in "