-
Notifications
You must be signed in to change notification settings - Fork 0
/
TODO
24 lines (19 loc) · 837 Bytes
/
TODO
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
High:
* append store scanner
* append store: compaction
* snapshot_delete: use snapshot bloom filter to delete data
* append_store_delete: scan append store to delete data
* log config
Medium:
* look append store cache
* local file system interface
Low:
* Reading configuration from file.
Done:
* testcases and examples : a) qfs test: file create, r/w/append. delete. b) append store tests
* append_store: sequential access for metadata
* snapshot write need to accumulate meta data write requests
* bug to fix: qfs doesn't allow r/w happen on the same chunk
1) append store need to switch between read and write mode
a) on read, check if the read chunk is the same as write chunk, if yes, need to let writer abdicate file ownership (call Close)
b) on append, check if opened, if not, close reader if there is one, then open.