I would like to understand how vLockStep is working and could only find very high level description.

The current documentation say that disk read but not disk write performed by the primary are replicated to the secondary. What is the reason for that ?

Also non-deterministict cpu instruction are also logged and replicated. I can only assume it's to cause the execution on the secondary to not execute those "non-deterministic instruction" and instead sync the effect of calling it from primary to secondary.

If this is the case, this would mean the Primary need to wait for the secondary to finish (executing/syncing) the non-deterministic instruction before sending the response to a network request. This seem like this would add significant latency.

Am I completely wrong in my understanding, or is this more or less correct request latency on the primary would be significantly slower than when not using vLockstep?

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.