Adventures in hugepages, part duex
Having solved my previous problem with hugepages, I move on to the next . . .
I have a little test server running under VBox. Configured hugepages. as before.
Problem is that when the server started, even though I specified vm.nr_hugepages = 1182, it only allocated 817. Further, when I tried to start a database, a process that normally takes about about 20 seconds, now with hugepages takes over 10 minutes.
The alert log for the database startup had several instances of background process timing out for failure to start in the expected time.
Immediately upon issuing the startup for the database the following message appeared in /var/log/messages: