Hi,
we started now monitoring the undo space after having a problem with a hanging Local checkpoint (LCP)
( https://forums.mysql.com/read.php?25,661125,662601#msg-662601 ).
A LCP takes about 2 1/2 hours to complete. The undo space usage goes down with the successful LCP - but it does not go down to 0% (see https://data.boerse-go.de/ndb-cluster-undo-space-usage.png ).
My expectancy is that after a LCP the undo space usage should become 0% again.
Is that not so?
Best regards
we started now monitoring the undo space after having a problem with a hanging Local checkpoint (LCP)
( https://forums.mysql.com/read.php?25,661125,662601#msg-662601 ).
A LCP takes about 2 1/2 hours to complete. The undo space usage goes down with the successful LCP - but it does not go down to 0% (see https://data.boerse-go.de/ndb-cluster-undo-space-usage.png ).
My expectancy is that after a LCP the undo space usage should become 0% again.
Is that not so?
Best regards