From: Sage Weil <sage@newdream.net>
Date: Thu, 6 May 2010 18:30:51 +0000 (-0700)
Subject: osd: clear pg REPLAY when we reinitiate peering
X-Git-Tag: v0.20.1~14
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=92a331473fdfcc684a8d4d99e99c233aa8b265b3;p=ceph.git

osd: clear pg REPLAY when we reinitiate peering

Among other things, a stray/bad REPLAY flag can trigger a false activation
like so:

osd/PG.cc: In function 'void PG::activate(ObjectStore::Transaction&, std::list<Context*, std::allocator<Context*> >&, std::map<int, MOSDPGInfo*, std::less<int>, std::allocator<std::pair<const int, MOSDPGInfo*> > >*)':
osd/PG.cc:1641: FAILED assert(peer_info.count(peer))
 1: (PG::activate(ObjectStore::Transaction&, std::list<Context*, std::allocator<Context*> >&, std::map<int, MOSDPGInfo*, std::less<int>, std::allocator<std::pair<int const, MOSDPGInfo*> > >*)+0x817) [0x75e637]
 2: (OSD::activate_pg(pg_t, utime_t)+0x136) [0x6bfe86]
 3: (OSD::check_replay_queue()+0x134) [0x6c0074]
 4: (OSD::tick()+0x1db) [0x6d96eb]
 5: (OSD::C_Tick::finish(int)+0x1c) [0x74844e]
 6: (SafeTimer::EventWrapper::finish(int)+0x6d) [0x7fa2cf]
 7: (Timer::timer_entry()+0x454) [0x7fb52a]
 8: (Timer::TimerThread::entry()+0x19) [0x647c09]
 9: (Thread::_entry_func(void*)+0x20) [0x659f18]
 10: /lib/libpthread.so.0 [0x7fceb73d473a]
 11: (clone()+0x6d) [0x7fceb65fe69d]
---

diff --git a/src/osd/OSD.cc b/src/osd/OSD.cc
index cb317ef3283..fc299896449 100644
--- a/src/osd/OSD.cc
+++ b/src/osd/OSD.cc
@@ -2399,6 +2399,7 @@ void OSD::advance_map(ObjectStore::Transaction& t)
     pg->state_clear(PG_STATE_DOWN);
     pg->state_clear(PG_STATE_PEERING);  // we'll need to restart peering
     pg->state_clear(PG_STATE_DEGRADED);
+    pg->state_clear(PG_STATE_REPLAY);
 
     if (pg->is_primary()) {
       if (osdmap->get_pg_size(pg->info.pgid) != pg->acting.size())
@@ -2476,7 +2477,6 @@ void OSD::advance_map(ObjectStore::Transaction& t)
 	if (role == 0) {
 	  // i am (still) primary. but my replica set changed.
 	  pg->state_clear(PG_STATE_CLEAN);
-	  pg->state_clear(PG_STATE_REPLAY);
 	  
 	  dout(10) << *pg << " " << oldacting << " -> " << pg->acting
 		   << ", replicas changed" << dendl;
diff --git a/src/osd/PG.h b/src/osd/PG.h
index 58304f2d7ee..77d0d8c7341 100644
--- a/src/osd/PG.h
+++ b/src/osd/PG.h
@@ -1103,8 +1103,8 @@ inline ostream& operator<<(ostream& out, const PG& pg)
     if (lost)
       out << " l=" << lost;
   }
-  if (pg.info.snap_trimq.size())
-    out << " snaptrimq=" << pg.info.snap_trimq;
+  //if (pg.info.snap_trimq.size())
+  //out << " snaptrimq=" << pg.info.snap_trimq;
 
   if (pg.deleting)
     out << " DELETING";