diff options
| -rw-r--r-- | src/common/config_opts.h | 2 | ||||
| -rw-r--r-- | src/osd/PG.cc | 1 | ||||
| -rw-r--r-- | src/osd/PGLog.h | 18 | ||||
| -rw-r--r-- | src/tools/ceph_objectstore_tool.cc | 3 |
4 files changed, 22 insertions, 2 deletions
diff --git a/src/common/config_opts.h b/src/common/config_opts.h index 23aaa6a71cf..723eebcc144 100644 --- a/src/common/config_opts.h +++ b/src/common/config_opts.h @@ -751,6 +751,8 @@ OPTION(osd_op_num_shards, OPT_INT, 5) OPTION(osd_op_queue, OPT_STR, "wpq") // PrioritzedQueue (prio), Weighted Priority Queue (wpq), or debug_random OPTION(osd_op_queue_cut_off, OPT_STR, "low") // Min priority to go to strict queue. (low, high, debug_random) +OPTION(osd_ignore_stale_divergent_priors, OPT_BOOL, false) // do not assert on divergent_prior entries which aren't in the log and whose on-disk objects are newer + // Set to true for testing. Users should NOT set this. // If set to true even after reading enough shards to // decode the object, any error will be reported. diff --git a/src/osd/PG.cc b/src/osd/PG.cc index aabb1290a72..f8c14131f6d 100644 --- a/src/osd/PG.cc +++ b/src/osd/PG.cc @@ -3289,6 +3289,7 @@ void PG::read_state(ObjectStore *store, bufferlist &bl) ghobject_t(info_struct_v < 8 ? OSD::make_pg_log_oid(pg_id) : pgmeta_oid), info, oss, + cct->_conf->osd_ignore_stale_divergent_priors, cct->_conf->osd_debug_verify_missing_on_start); if (oss.tellp()) osd->clog->error() << oss.rdbuf(); diff --git a/src/osd/PGLog.h b/src/osd/PGLog.h index 5096e203179..e4d0fa304ab 100644 --- a/src/osd/PGLog.h +++ b/src/osd/PGLog.h @@ -1098,11 +1098,13 @@ public: coll_t log_coll, ghobject_t log_oid, const pg_info_t &info, ostringstream &oss, + bool tolerate_divergent_missing_log, bool debug_verify_stored_missing = false ) { return read_log_and_missing( store, pg_coll, log_coll, log_oid, info, log, missing, oss, + tolerate_divergent_missing_log, &clear_divergent_priors, this, (pg_log_debug ? &log_keys_debug : 0), @@ -1115,6 +1117,7 @@ public: const pg_info_t &info, IndexedLog &log, missing_type &missing, ostringstream &oss, + bool tolerate_divergent_missing_log, bool *clear_divergent_priors = NULL, const DoutPrefixProvider *dpp = NULL, set<string> *log_keys_debug = 0, @@ -1290,7 +1293,20 @@ public: * version would not have been recovered, and a newer version * would show up in the log above. */ - assert(oi.version == i->first); + /** + * Unfortunately the assessment above is incorrect because of + * http://tracker.ceph.com/issues/17916 (we were incorrectly + * not removing the divergent_priors set from disk state!), + * so let's check that. + */ + if (oi.version > i->first && tolerate_divergent_missing_log) { + ldpp_dout(dpp, 0) << "read_log divergent_priors entry (" << *i + << ") inconsistent with disk state (" << oi + << "), assuming it is tracker.ceph.com/issues/17916" + << dendl; + } else { + assert(oi.version == i->first); + } } else { ldpp_dout(dpp, 15) << "read_log_and_missing missing " << *i << dendl; missing.add(i->second, i->first, eversion_t()); diff --git a/src/tools/ceph_objectstore_tool.cc b/src/tools/ceph_objectstore_tool.cc index f9bf85a5d63..c4897805c38 100644 --- a/src/tools/ceph_objectstore_tool.cc +++ b/src/tools/ceph_objectstore_tool.cc @@ -328,7 +328,8 @@ int get_log(ObjectStore *fs, __u8 struct_ver, PGLog::read_log_and_missing(fs, coll, struct_ver >= 8 ? coll : coll_t::meta(), struct_ver >= 8 ? pgid.make_pgmeta_oid() : log_oid, - info, log, missing, oss); + info, log, missing, oss, + g_ceph_context->_conf->osd_ignore_stale_divergent_priors); if (debug && oss.str().size()) cerr << oss.str() << std::endl; } |