How to ignore or replace duplicate data in clickhouse?

Question

SQL

Good day.
Actually the question is simple.
Table snippet

...
`id` String,
`install_time` DateTime
)
ENGINE = ReplacingMergeTree
ORDER BY (id, install_time)

insert query is the simplest
INSERT INTO installs (f1,f2) values ​​(v1,v2)

but the records are suddenly duplicated.

and the query OPTIMIZE TABLE installs DEDUPLICATE;
works fine and duplicates disappear.
Does it really need to be run every time?

Or is there a way to teach the table to skip or replace duplicates?