This morning a colleague of mine noticed an alert from our plan change monitor and prevented any impact on the users by forcing a SQL’s plan to its normal fast plan before the users could be affected. I want to share the monitor script and describe how we use its output.
I know this is long but I want to paste the SQL text of the monitor script here. Review my comments on each step. I can not remember the details of the script myself even though I wrote it but the comments hopefully will help:
I’m genuinely delighted that Graham Smith has added his profile to our ‘How I Work‘ series. Not just because he’s one of the World’s best known PeopleSoft experts, or just because he’s recently joined Cedar’s family of PeopleSoft consultants, but mainly because I was sure he’d give some really interesting answers.
There can be a lot of confusion and lack of planning in Percona XtraDB Clusters in regards to nodes becoming desynchronized for various reasons. This can happen a few ways: