Data to gather for "long semaphore wait"


I have a customer who saw a bunch of errors: long semaphore wait. He seems to have seen a bunch of them over a couple of days. What information should I collect to help determine what may be causing this error? Would I be able to setup pt-stalk to automatically trigger and collect data if this happens again?



Semaphore waits can happen for a number of reasons. Setting up pt-stalk is a good idea. Setting up PMM would be even better to see exactly what is going on at the time of the incident.

1 Like

@Ivan_Groenewold Thank-you for the information. How would I set up PMM so that I can collect data when this happens again?

1 Like

Check out this page to get started: Install Percona Monitoring and Management

1 Like