I’m working on couchbase clusters of our client with c#sdk. During the develop phase we’ve switch from durability none to durability majorityAndPersistActive, due to our business data consistency. But once set this level and timeout at 10 s (reached after several tries staring from 2 5), we are receiving several exception couchbase.core.exceptions.ambiguoustimeoutexception on upsert. The server is latest version and also the sdk. Do you retain useful to ask the client an investigation about poor performance? Is it acceptable to retain correct to get no error with 10 s of timeout?
We have the similar case with durability level of majorityAndPersistActive. I couldn’t find an answer but i can say that durable writes are unperformant ones comparing none durable writes or durable with just majority ones. We have seen huge difference while persisting to the disk.
Is this with a single node cluster or with a cluster with multiple nodes? Again as I stated in the other post there are big differences between running on docker with a single node and running multiple nodes. As stated in our docs many features do not work properly on a single node cluster:
In the other post, you never answered if adding multiple nodes fixed your issue or not. If it didn’t, we would want to get logs over to the SDK to review.