Github user paul-rogers commented on the issue:
https://github.com/apache/drill/pull/846
Another critical issue here. Direct memory is not a limitless resource,
unfortunately. Allocating memory larger than 16 MB causes memory fragmentation
as the allocation must come from the system, but all free memory is cached in
Netty in 16 MB chunks. So, if the buffer size here is larger than 16 MB, the
operation may fail with an OOM.
If this operation must buffer in memory, and must use larger than 16 MB,
then you need something like what the original code provided. That buffered
writer can be backed by a chain of 16 MB direct memory blocks.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
