An unsequence file that covers too many sequence file causes OOM query

[qiaojialin]

version 0.10:

Due to the separation of sequence and unsequenced data, the unsequenced file may cover a long time range, which overlaps too many sequence files.

For example,
the time range of sequence file is [1,10], [13,20], [21,30], [31,40],..., [1000,1010]
unsequence file is [12, 1000].

If we query [11, 1000], the unsequence file is the first file in the query range, and we read all its overlapped Chunks into memory. Then OOM occurs.