Thanks Boris, that is interesting as disabling PFS in that way is still supposed to leave the performance_schema.session_variables and performance_schema.global_variables tables operational and I have validated as much. I wonder if maybe there is some path through upgrade where these tables don't get some initial creation/activation. If so, then there may be some upstream issue. This behavior is documented here https://dev.mysql.com/doc/refman/5.7/en/performance-schema-system-variables.html#sysvar_performance_schema
So to repeat what I said before, I can not find a way to disable these tables unless I actually compiled it out, which no standard distribution or package does that I know of. Since there are a few of you hitting it, there must be some path where these tables remain inaccessible or inactive.
Thanks Boris, that is interesting as disabling PFS in that way is still supposed to leave the performance_ schema. session_ variables and performance_ schema. global_ variables tables operational and I have validated as much. I wonder if maybe there is some path through upgrade where these tables don't get some initial creation/ activation. If so, then there may be some upstream issue. This behavior is documented here https:/ /dev.mysql. com/doc/ refman/ 5.7/en/ performance- schema- system- variables. html#sysvar_ performance_ schema
So to repeat what I said before, I can not find a way to disable these tables unless I actually compiled it out, which no standard distribution or package does that I know of. Since there are a few of you hitting it, there must be some path where these tables remain inaccessible or inactive.