From 0d95a0db3ec1af3cbf7ff3174dc677714568e81c Mon Sep 17 00:00:00 2001
From: Sage Weil <sage@newdream.net>
Date: Thu, 31 Jan 2008 12:01:31 -0800
Subject: [PATCH] mds: dont force journal flush on every op; mdlog+journaler
 are smart enough to detect waiters

---
 src/TODO       | 38 ++++++--------------------------------
 src/mds/MDS.cc |  2 +-
 2 files changed, 7 insertions(+), 33 deletions(-)

diff --git a/src/TODO b/src/TODO
index 7b3c0740546c6..72f90388a401e 100644
--- a/src/TODO
+++ b/src/TODO
@@ -17,8 +17,6 @@ code cleanup
   - probably kill base case in encoder.h, replace with int types, with appropriate swabbing
   - specify word size in all encoded structures
 
-
-
 kernel client
 - some bugs
  - open path needs to call fill_trace/fill_inode such that ci->i_max_size gets filled in
@@ -125,9 +123,7 @@ mds
   - and dir sizes, if i add that... also on rejoin?
 
 - efficient stat for single writers
-- lstat vs stat?
 - add FILE_CAP_EXTEND capability bit
-- delayed replica caps release... we need to set a timer event? (and cancel it when appropriate?)
 
 
 client
@@ -150,6 +146,11 @@ journaler
 - need to truncate at detected (valid) write_pos to clear out any other partial trailing writes
 
 
+fsck
+- fsck.ebofs
+- online mds fsck?
+
+
 
 
 rados snapshots
@@ -222,6 +223,7 @@ messenger
 - fix messenger shutdown.. we shouldn't delete messenger, since the caller may be referencing it, etc.
 
 simplemessenger
+- fix/audit accept() logic to detect reset, do callback
 - close idle connections
 - take a look at RDS?  http://oss.oracle.com/projects/rds/
 
@@ -287,31 +289,3 @@ ebofs
 remaining hard problems
 - how to cope with file size changes and read/write sharing
 
-
-
-
-
-
-
-
-
-
-
-
-
-why qsync could be wrong (for very strict POSIX) : varying mds -> client message transit or processing times.
-- mds -> 1,2 : qsync
-- client1 writes at byte 100
-- client1 -> mds : qsync reply (size=100)
-- client1 writes at byte 300
-- client1 -> client2 (outside channel)
-- client2 writes at byte 200
-- client2 -> mds : qsync reply (size=200)
--> stat results in size 200, even though at no single point in time was the max size 500.
--> for correct result, need to _stop_ client writers while gathering metadata.
-
-
-- dump active config in run output somewhere
-
-
-
diff --git a/src/mds/MDS.cc b/src/mds/MDS.cc
index 1b8448432a599..418d520d16f59 100644
--- a/src/mds/MDS.cc
+++ b/src/mds/MDS.cc
@@ -1136,7 +1136,7 @@ void MDS::_dispatch(Message *m)
   // HACK FOR NOW
   if (is_active() || is_stopping()) {
     // flush log to disk after every op.  for now.
-    mdlog->flush();
+    //mdlog->flush();
 
     // trim cache
     mdcache->trim();
-- 
2.39.5