Beispiel #1
0
    // @param reconf true if this is a reconfiguration and not an initial load of the configuration.
    // @return true if ok; throws if config really bad; false if config doesn't include self
    bool ReplSetImpl::initFromConfig(OperationContext* txn, ReplSetConfig& c, bool reconf) {
        // NOTE: haveNewConfig() writes the new config to disk before we get here.  So
        //       we cannot error out at this point, except fatally.  Check errors earlier.
        lock lk(this);

        if (!getLastErrorDefault.isEmpty() || !c.getLastErrorDefaults.isEmpty()) {
            getLastErrorDefault = c.getLastErrorDefaults;
        }

        list<ReplSetConfig::MemberCfg*> newOnes;
        // additive short-cuts the new config setup. If we are just adding a
        // node/nodes and nothing else is changing, this is additive. If it's
        // not a reconfig, we're not adding anything
        bool additive = reconf;
        bool updateConfigs = false;
        {
            unsigned nfound = 0;
            int me = 0;
            for (vector<ReplSetConfig::MemberCfg>::iterator i = c.members.begin();
                    i != c.members.end();
                    i++) {
                
                ReplSetConfig::MemberCfg& m = *i;
                if (isSelf(m.h)) {
                    me++;
                }
                
                if (reconf) {
                    const Member *old = findById(m._id);
                    if (old) {
                        nfound++;
                        verify((int) old->id() == m._id);
                        if (!old->config().isSameIgnoringTags(m)) {
                            additive = false;
                        }
                        if (!updateConfigs && old->config() != m) {
                            updateConfigs = true;
                        }
                    }
                    else {
                        newOnes.push_back(&m);
                    }
                }
            }
            if (me == 0) { // we're not in the config -- we must have been removed
                if (state().removed()) {
                    // already took note of our ejection from the set
                    // so just sit tight and poll again
                    return false;
                }

                _members.orphanAll();

                // kill off rsHealthPoll threads (because they Know Too Much about our past)
                endOldHealthTasks();

                // clear sync target to avoid faulty sync attempts; we must do this before we
                // close sockets, since that will trigger the bgsync thread to reconnect.
                BackgroundSync::get()->clearSyncTarget();

                // close sockets to force clients to re-evaluate this member
                MessagingPort::closeAllSockets(0);

                // take note of our ejection
                changeState(MemberState::RS_REMOVED);

                // go into holding pattern
                log() << "replSet info self not present in the repl set configuration:" << rsLog;
                log() << c.toString() << rsLog;

                loadConfig(txn);  // redo config from scratch
                return false; 
            }
            uassert(13302, "replSet error self appears twice in the repl set configuration", me<=1);

            if (state().removed()) {
                // If we were removed and have now been added back in, switch state.
                changeState(MemberState::RS_RECOVERING);
            }

            // if we found different members that the original config, reload everything
            if (reconf && config().members.size() != nfound)
                additive = false;
        }

        // If we are changing chaining rules, we don't want this to be an additive reconfig so that
        // the primary can step down and the sync targets change.
        // TODO: This can be removed once SERVER-5208 is fixed.
        if (reconf && config().chainingAllowed() != c.chainingAllowed()) {
            additive = false;
        }

        _cfg = new ReplSetConfig(c);

        // config() is same thing but const, so we use that when we can for clarity below
        dassert(&config() == _cfg);
        verify(config().ok());
        verify(_name.empty() || _name == config()._id);
        _name = config()._id;
        verify(!_name.empty());

        // this is a shortcut for simple changes
        if (additive) {
            log() << "replSet info : additive change to configuration" << rsLog;
            if (updateConfigs) {
                // we have new configs for existing members, so we need to repopulate _members
                // with the most recent configs
                _members.orphanAll();

                // for logging
                string members = "";

                // not setting _self to 0 as other threads use _self w/o locking
                int me = 0;
                for(vector<ReplSetConfig::MemberCfg>::const_iterator i = config().members.begin();
                    i != config().members.end(); i++) {
                    const ReplSetConfig::MemberCfg& m = *i;
                    Member *mi;
                    members += (members == "" ? "" : ", ") + m.h.toString();
                    if (isSelf(m.h)) {
                        verify(me++ == 0);
                        mi = new Member(m.h, m._id, &m, true);
                        setSelfTo(mi);
                    }
                    else {
                        mi = new Member(m.h, m._id, &m, false);
                        _members.push(mi);
                    }
                }
                // trigger a handshake to update the syncSource of our writeconcern information
                syncSourceFeedback.forwardSlaveHandshake();
            }

            // add any new members
            for (list<ReplSetConfig::MemberCfg*>::const_iterator i = newOnes.begin();
                    i != newOnes.end();
                    i++) {
                ReplSetConfig::MemberCfg *m = *i;
                Member *mi = new Member(m->h, m->_id, m, false);

                // we will indicate that new members are up() initially so that we don't relinquish
                // our primary state because we can't (transiently) see a majority. they should be
                // up as we check that new members are up before getting here on reconfig anyway.
                mi->get_hbinfo().health = 0.1;

                _members.push(mi);
                startHealthTaskFor(mi);
            }

            // if we aren't creating new members, we may have to update the
            // groups for the current ones
            _cfg->updateMembers(_members);

            return true;
        }

        // start with no members.  if this is a reconfig, drop the old ones.
        _members.orphanAll();

        endOldHealthTasks();
        
        int oldPrimaryId = -1;
        {
            const Member *p = box.getPrimary();
            if (p)
                oldPrimaryId = p->id();
        }
        forgetPrimary(txn);

        // not setting _self to 0 as other threads use _self w/o locking
        int me = 0;

        // For logging
        string members = "";

        for (vector<ReplSetConfig::MemberCfg>::const_iterator i = config().members.begin();
                i != config().members.end();
                i++) {
            const ReplSetConfig::MemberCfg& m = *i;
            Member *mi;
            members += (members == "" ? "" : ", ") + m.h.toString();
            if (isSelf(m.h)) {
                verify(me++ == 0);
                mi = new Member(m.h, m._id, &m, true);
                if (!reconf) {
                    log() << "replSet I am " << m.h.toString() << rsLog;
                }
                setSelfTo(mi);

                if ((int)mi->id() == oldPrimaryId)
                    box.setSelfPrimary(mi);
            }
            else {
                mi = new Member(m.h, m._id, &m, false);
                _members.push(mi);
                if ((int)mi->id() == oldPrimaryId)
                    box.setOtherPrimary(mi);
            }
        }

        if (me == 0){
            log() << "replSet warning did not detect own host in full reconfig, members "
                  << members << " config: " << c << rsLog;
        }
        else {
            // Do this after we've found ourselves, since _self needs
            // to be set before we can start the heartbeat tasks
            for (Member *mb = _members.head(); mb; mb=mb->next()) {
                startHealthTaskFor(mb);
            }
        }
        return true;
    }