T
TanStack•17h ago
stormy-gold

Electric Replica Identity

electric-1  | 00:32:34.930 pid=<0.3276.0> [error] Schema reconciliation failed: ** (Electric.DbConfigurationError) Database table "public.users" does not have its replica identity set to FULL
electric-1  | 00:32:34.930 pid=<0.3276.0> [error] Schema reconciliation failed: ** (Electric.DbConfigurationError) Database table "public.users" does not have its replica identity set to FULL
just wondering if there are plans to support non-full (ie default) replica identities in the future? or is this requirement by design?
3 Replies
eager-peach
eager-peach•17h ago
FULL is necessary to handle a number of different scenarios where we need both the old and new values to correctly figure out what changed is FULL a problem for y'all?
stormy-gold
stormy-goldOP•16h ago
definitely not a problem at my current scale 😂 I was slightly nervous about it after reading the pg docs for replication, which say "replica identity FULL should only be used as a fallback if no other solution is possible", but doing some more reading (https://xata.io/blog/replica-identity-full-performance) makes it sound like in practice pg will use the table's PK anyways
Xata
On the performance impact of REPLICA IDENTITY FULL in Postgres by T...
Analyzing the performance impact of REPLICA IDENTITY FULL by looking at the PostgreSQL source code and performing benchmarks.
stormy-gold
stormy-goldOP•16h ago
thanks as usual for the quick responses!

Did you find this page helpful?