baritra

@porunov Thanks for this too, as we are also facing similar index staleness issues. However, how do we prevent it in the first place?

For your earlier description -- one possible prevention mechanism could be to minimize shutdowns during update -- by adding more graceful shutdowns. What are potential other prevention methods @porunov

Thanks again @porunov Does that mean I should set `storage.cql.atomic-batch-mutate` to `true` and `storage.cql.batch-statement-size` to a value that's greater than the maximum number of vertices mutated in my transactions?

Thanks a lot @porunov !! I will try those. I will also look at #statements for our largest transactions (not sure how to do that yet) We are unfortunately stuck...

@porunov I experimented with the batch size and batch atomicity properties, and had the following findings, and follow up questions. 1. It seems that if I turn on batch-atomicity, batch...

@porunov @li-boxuan -- will be really grateful if you could shed some light on the above.

@porunov @li-boxuan I think atomic batches are working completely fine. However, **I think there might be a bug with the unlogged (non-atomic) batching of CQL statements. Here's why.** For non-atomic...