The full warning is [Warning] InnoDB: Difficult to find free blocks in the buffer pool (21 search iterations)! 0 failed attempts to flush a page!
The four databases reporting this show the exact warning including the 21 search iterations.
We have increased the buffer pool size on these servers and are still getting the warnings
Buffer pool sizes on these 5 databases
innodb_buffer_pool_size=1024M #Set to 70% of available RAM
innodb_buffer_pool_size=2100M #Set to 70% of available RAM
innodb_buffer_pool_size=2048M #Set to 70% of available RAM
innodb_buffer_pool_size=2450M #Set to 70% of available RAM
innodb_buffer_pool_size=90000M #Set to 70% of available RAM
The first buffer pool size listed was the size of the buffer pool used during the upgrade and when we first detected the warning in the logs. The second buffer pool setting is what we increased the value to as the warning recommends...yet it did not resolve the warning on any of those databases that had an increase in buffer pool size.
After the upgrade whenever we restart the database we will get this error on some of the servers within 24 hours. A couple of databases have generated the warning more than a week after the last database restart. Some of these databases have many gigabytes of buffer_pool assigned.
Maybe related to https://jira.mariadb.org/browse/MDEV-16339
Please advise on the significance of this warning and how to resolve it.