todo: mds

diff --git a/src/TODO b/src/TODO
index 26564da1939295f3ed048012347533e1a086dfda..81827306e72d88b7a8229d76d911c103298fd15b 100644 (file)
--- a/src/TODO
+++ b/src/TODO
@@ -101,6 +101,7 @@ userspace client
 - fix readdir vs fragment race by keeping a separate frag pos, and ignoring dentries below it
 
 mds
+- file recovery maybe needs to scan entire file range for a truncation event?  (and object attr maybe needs file offset, not object offset, or original truncation?)
 - on replay, but dirty scatter replicas on lists so that they get flushed?  or does rejoin handle that?
 - take some care with replayed client requests vs new requests
 - linkage vs cdentry replicas and remote rename....