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]
This commit is contained in:
Sage Weil 2010-05-06 11:30:51 -07:00
parent 642be312ba
commit 66350d2b2b
2 changed files with 3 additions and 3 deletions

View File

@ -2400,6 +2400,7 @@ 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())
@ -2477,7 +2478,6 @@ 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;

View File

@ -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";