Hi everyone! Has anyone here dealt nodes being out of sync? We have 3 nodes that somehow managed to get out of sync: • querying is returning varying results with the

out_of

contents also varying significantly For example, a query is supposed to return 8 results (when we compare this to our master data set). When we query, here is an example of the varying results we will get: • 6 hits, with out_of: 228,557 • 8 hits, with out_of: 213,157 (this 8 hits are correct in this case, but the item count is significantly off) For context: we've been on TypeSense Cloud v28.0.rc12 and currently upgrading to v28.0.rc33. But irrespective of the version, I'm curious if anyone has faced issues with varying results from different nodes, and what they've done to mitigate this issue going forward. Any help is appreciated!

Thanks for the details about the node sync issues. Could you let us know how long you've noticed the nodes being out of sync? Understanding the timeline would help diagnose what might have triggered this. For context on replication: When data is written to your primary node, it starts a replication process where: 1. The primary node logs the change 2. It initiates replication streams to each secondary node 3. Each secondary node receives and applies these changes to match the primary If the searches happened as the first node got some documents indexed, that may explain why the other two nodes didn't have the same number of documents, as it takes some amount of time to index the documents on the other nodes as well