We have PXC 5.7 cluster with ~130 databases and ~100000 tables.
When we do mysqlbackup we see that AnonPages memory usage is going up and after mysqlbackup doesn’t go down.
In test environment we have 24GB RAM (innodb_buffer_pool_size=16G) and after we do multiple mysqldumps for all databases we got OOM.
It’s happening only on that node where we run mysqldump.
At the moment we have idea that maybe we need more memory, but we want to understand why, is that because we have so many tables or there is something else?
Maybe someone had issue with similar symptoms when you see that it’s AnonPages what goes only up?
Hi, have you tried option –quick that should reduce memory usage.
I will try to use --quick option, but I don’t think it will help, because the same issue is when we used xtrabackup, just mysqldump is easier to reproduce.
I was hoping maybe someone had similar issues when there is lot databases with lot of tables.
Update:
mysqldump --quick option doesn’t help at all
But, when we gave VM 64GB RAM and we set innodb_buffer_pool_size=8GB after few days of running backups we got AnonPages arounfd 22GB-25GB and it stopped growing.
I see that it’s based on how much tabes and data we have, but maybe someone has better explanation?