This commit adds a caveat to the README to notify users that bulk loading
random writes into a single new bucket can be problematic. This issue was
reported by @freepk.
Fixes#244.
This commit fixes a bug that occurs when a root node is split just after a re-mmap occurs.
Previously, this would cause a panic because the new root node would still reference
keys from the old mmap.
This commit moves the DB.FillPercent field to Bucket.FillPercent. This
allows the fill percentage to be specified per-bucket, per-tx. This
value is not persisted and should be set whenever using it.
This commit fixes a bug where a root split on a very large insert would
cause an overflow of the root node. The problem was that the new root
was not split when it was created so a new root with more than 64K child
nodes would overflow the page.count (uint16).
This fixes an issue where split nodes can be double spilled. This is typically
not noticable but it can have large effects when bulk inserting as double
spilled nodes will get added to the freelist which will grow quickly.
This commit fixes an issue where large nodes would take up most of
the insert time because the entire node size would be calculated to
check if it could fit in a page or not. This changes the behavior
so that a node's size is only calculated up to the size it needs to
check and then returns.
This commit adds the DB.NoSync flag to skip fsync() calls on each commit. This should only
be used for bulk loading as it can corrupt your database in the event of a system failure.
Initial tests show it can provide a 2x speed up for sequential inserts.
This commit changes TxStats.RebalanceTime to only update if there are nodes that have been
rebalanced. Previously, the RebalanceTime would include a small amount of time to check if
there were nodes to rebalance but this is confusing to users and not terribly helpful.
This commit adds a defer handler to ensure that transactions are always closed out - even
in the event of a panic within user code. It's recommended that applications always fail
in the event of a panic but some packages such as net/http will automatically recover
which is a problem (IHMO).