org.jgroups.protocols.pbcast.Merger Maven / Gradle / Ivy
package org.jgroups.protocols.pbcast;
import org.jgroups.*;
import org.jgroups.logging.Log;
import org.jgroups.util.*;
import java.util.*;
import java.util.concurrent.ConcurrentMap;
import java.util.concurrent.Future;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.TimeoutException;
import static org.jgroups.Message.Flag.OOB;
import static org.jgroups.Message.TransientFlag.DONT_LOOPBACK;
/**
* Handles merging. Called by CoordGmsImpl and ParticipantGmsImpl
* @author Bela Ban
*/
public class Merger {
protected final GMS gms;
protected final Log log;
protected final MergeTask merge_task=new MergeTask();
/** For MERGE_REQ/MERGE_RSP correlation, contains MergeData elements */
protected final ResponseCollector merge_rsps=new ResponseCollector<>();
/** For GET_DIGEST / DIGEST_RSP correlation */
protected final ResponseCollector digest_collector=new ResponseCollector<>();
protected MergeId merge_id=null;
protected final BoundedList merge_id_history=new BoundedList<>(20);
protected Future merge_killer=null;
public Merger(GMS gms) {
this.gms=gms;
this.log=gms.getLog();
}
public String getMergeIdAsString() {return merge_id != null? merge_id.toString() : null;}
public String getMergeIdHistory() {return merge_id_history.toString();}
public boolean isMergeTaskRunning() {return merge_task.isRunning();}
public boolean isMergeKillerTaskRunning() {return merge_killer != null && !merge_killer.isDone();}
public synchronized MergeId getMergeId() {return merge_id;} // only used for testing; do not use
public synchronized boolean isMergeInProgress() {return merge_id != null;}
public synchronized boolean matchMergeId(MergeId id) {return Util.match(this.merge_id, id);}
public synchronized boolean setMergeId(MergeId expected, MergeId new_value) {
boolean match=Util.match(this.merge_id, expected);
if(match) {
if(new_value != null && merge_id_history.contains(new_value))
return false;
else
merge_id_history.add(new_value);
this.merge_id=new_value;
if(this.merge_id != null) {
// Clears the view handler queue and discards all JOIN/LEAVE/MERGE requests until after the MERGE
gms.getViewHandler().suspend();
gms.getDownProtocol().down(new Event(Event.SUSPEND_STABLE, 20000));
startMergeKiller();
}
}
return match;
}
/**
* Invoked upon receiving a MERGE event from the MERGE layer. Starts the merge protocol.
* See description of protocol in DESIGN.
* @param views A List of different views detected by the merge protocol, keyed by sender
*/
public void merge(Map views) {
if(views == null || views.isEmpty()) {
log.warn("the views passed with the MERGE event were empty (or null); ignoring MERGE event");
return;
}
if(View.sameViews(views.values())) {
log.debug("MERGE event is ignored because of identical views: %s", Util.printListWithDelimiter(views.values(), ", "));
return;
}
if(isMergeInProgress()) {
log.trace("%s: merge is already running (merge_id=%s)", gms.getAddress(), merge_id);
return;
}
Address merge_leader=determineMergeLeader(views);
if(merge_leader == null)
return;
if(merge_leader.equals(gms.getAddress())) {
log.debug("%s: I will be the merge leader. Starting the merge task. Views: %s", gms.getAddress(), views);
merge_task.start(views);
}
else
log.trace("%s: I'm not the merge leader, waiting for merge leader (%s) to start merge", gms.getAddress(), merge_leader);
}
/**
* Get the view and digest and send back both (MergeData) in the form of a MERGE_RSP to the sender.
* If a merge is already in progress, send back a MergeData with the merge_rejected field set to true.
* @param sender The address of the merge leader
* @param merge_id The merge ID
* @param mbrs The set of members from which we expect responses. Guaranteed to be non-null
*/
public void handleMergeRequest(Address sender, MergeId merge_id, Collection mbrs) {
try {
_handleMergeRequest(sender, merge_id, mbrs);
}
catch(Throwable t) {
log.error("%s: failure handling the merge request: %s", gms.getAddress(), t.getMessage());
cancelMerge(merge_id);
sendMergeRejectedResponse(sender, merge_id);
}
}
public void handleMergeResponse(MergeData data, MergeId merge_id) {
if(!matchMergeId(merge_id)) {
log.trace("%s: this.merge_id (%s) is different from merge_id %s sent by %s as merge response, discarding it",
gms.getAddress(), this.merge_id, merge_id, data.getSender());
return;
}
merge_rsps.add(data.getSender(), data);
}
/**
* If merge_id is not equal to this.merge_id then discard. Else cast the view/digest to all members of this group.
*/
public void handleMergeView(final MergeData data, final MergeId merge_id) {
if(!matchMergeId(merge_id)) {
log.trace("%s: merge_ids (mine: %s, received: %s) don't match; merge view %s is discarded",
gms.getAddress(), this.merge_id, merge_id, data.view.getViewId());
return;
}
// only send to our *current* members, if we have A and B being merged (we are B), then we would *not*
// want to block on a VIEW_ACK from A because A doesn't see us in the pre-merge view yet and discards the view
List expected_acks=gms.members.getMembers();
try {
gms.castViewChangeAndSendJoinRsps(data.view, data.digest, expected_acks, null, null);
// if we have flush in stack send ack back to merge coordinator
if(gms.flushProtocolInStack) { //[JGRP-700] - FLUSH: flushing should span merge
Message ack=new EmptyMessage(data.getSender()).setFlag(OOB)
.putHeader(gms.getId(), new GMS.GmsHeader(GMS.GmsHeader.INSTALL_MERGE_VIEW_OK));
gms.getDownProtocol().down(ack);
}
}
finally {
cancelMerge(merge_id);
}
}
public void handleMergeCancelled(MergeId merge_id) {
try {
gms.stopFlush();
}
catch(Throwable t) {
log.error(Util.getMessage("StopFlushFailed"), t.getMessage());
}
log.trace("%s: merge %s is cancelled", gms.getAddress(), merge_id);
cancelMerge(merge_id);
}
public void handleDigestResponse(Address sender, Digest digest) {
digest_collector.add(sender, digest);
}
/**
* Removes all members from a given view which don't have us in their view
* (https://issues.redhat.com/browse/JGRP-1061). Example:
*
* A: AB
* B: AB
* C: ABC
*
* becomes
*
* A: AB
* B: AB
* C: C // A and B don't have C in their views
*
* @param map A map of members and their associated views
*/
public static void sanitizeViews(Map map) {
if(map == null)
return;
for(Map.Entry entry: map.entrySet()) {
Address key=entry.getKey();
List members=new ArrayList<>(entry.getValue().getMembers());
boolean modified=false;
for(Iterator it=members.iterator(); it.hasNext();) {
Address val=it.next();
if(val.equals(key)) // we can always talk to ourself !
continue;
View view=map.get(val);
final Collection tmp_mbrs=view != null? view.getMembers() : null;
if(tmp_mbrs != null && !tmp_mbrs.contains(key)) {
it.remove();
modified=true;
}
}
if(modified) {
View old_view=entry.getValue();
entry.setValue(new View(old_view.getViewId(), members));
}
}
}
/** Returns the address of the merge leader */
protected Address determineMergeLeader(Map views) {
// we need the merge *coordinators* not merge participants because not everyone can lead a merge !
Collection coords=Util.determineActualMergeCoords(views);
if(coords.isEmpty())
coords=Util.determineMergeCoords(views); // https://issues.redhat.com/browse/JGRP-2092
if(coords.isEmpty()) {
log.error("%s: unable to determine merge leader from %s; not starting a merge", gms.getAddress(), views);
return null;
}
return new Membership(coords).sort().elementAt(0); // establish a deterministic order, so that coords can elect leader
}
/**
* Needs to return a map of all subview coordinators and their views (as a collection of members). The merge policy
* is defined in https://issues.redhat.com/browse/JGRP-1910
*/
protected static Map> determineMergeCoords(Map views) {
Map> retval=new HashMap<>();
for(View view: views.values()) {
Address coord=view.getCreator();
Collection members=retval.computeIfAbsent(coord, k -> new ArrayList<>());
for(Address mbr: view.getMembersRaw())
if(!members.contains(mbr))
members.add(mbr);
}
// For the merge participants which are not coordinator, we simply add them, and the associated
// membership list consists only of themselves
Collection merge_participants=Util.determineMergeParticipants(views);
merge_participants.removeAll(retval.keySet());
merge_participants.stream().filter(merge_participant -> !retval.containsKey(merge_participant))
.forEach(merge_participant -> retval.put(merge_participant, Collections.singletonList(merge_participant)));
return retval;
}
protected void _handleMergeRequest(Address sender, MergeId merge_id, Collection mbrs) throws Exception {
MergeId current_merge_id=this.merge_id;
boolean success=matchMergeId(merge_id) || setMergeId(null, merge_id);
if(!success) {
log.trace("%s: merge %s is already in progress, received merge-id=%s", gms.getAddress(), current_merge_id, merge_id);
return;
}
/* Clears the view handler queue and discards all JOIN/LEAVE/MERGE requests until after the MERGE */
// gms.getViewHandler().suspend();
log.trace("%s: got merge request from %s, merge_id=%s, mbrs=%s", gms.getAddress(), sender, merge_id, mbrs);
// merge the membership of the current view with mbrs
List members=new ArrayList<>(mbrs != null? mbrs.size() : 32);
if(mbrs != null) // didn't use a set because we didn't want to change the membership order at this time (although not incorrect)
mbrs.stream().filter(mbr -> !members.contains(mbr)).forEach(members::add);
// Now remove members that are not in our current view:
members.retainAll(gms.view().getMembers());
// Add myself if not present
if(!members.contains(gms.getAddress()))
members.add(gms.getAddress());
ViewId tmp_vid=gms.getViewId();
if(tmp_vid == null)
throw new Exception("view ID is null; cannot return merge response");
View view=new View(tmp_vid, members);
if(gms.flushProtocolInStack && !gms.startFlush(view)) // if flush is in stack, let this coord flush its subcluster
throw new Exception("flush failed");
// we still need to fetch digests from all members, and not just return our own digest (https://issues.redhat.com/browse/JGRP-948)
Digest digest=fetchDigestsFromAllMembersInSubPartition(view, merge_id);
if(digest == null || digest.capacity() == 0)
throw new Exception("failed fetching digests from subpartition members; dropping merge response");
sendMergeResponse(sender, view, digest, merge_id);
}
/** Send back a response containing view and digest to sender */
protected void sendMergeResponse(Address sender, View view, Digest digest, MergeId merge_id) {
Message msg=new BytesMessage(sender).setArray(GMS.marshal(view, digest)).setFlag(OOB)
.putHeader(gms.getId(), new GMS.GmsHeader(GMS.GmsHeader.MERGE_RSP).mergeId(merge_id));
gms.getDownProtocol().down(msg);
}
/**
* Sends the new view and digest to all subgroup coordinators. Each coord will in turn broadcast the new view and
* digest to all the members of its subgroup
*/
protected void sendMergeView(Collection coords, MergeData combined_merge_data, MergeId merge_id) {
if(coords == null || coords.isEmpty() || combined_merge_data == null)
return;
View view=combined_merge_data.view;
Digest digest=combined_merge_data.digest;
if(view == null || digest == null) {
log.error(Util.getMessage("ViewOrDigestIsNullCannotSendConsolidatedMergeView/Digest"));
return;
}
int size=0;
if(gms.flushProtocolInStack) {
gms.merge_ack_collector.reset(coords);
size=gms.merge_ack_collector.size();
}
Event install_merge_view_evt=new Event(Event.INSTALL_MERGE_VIEW, view);
gms.getUpProtocol().up(install_merge_view_evt);
gms.getDownProtocol().down(install_merge_view_evt);
// Change coords to move self to the tail: if UNICAST3.loopback is true, we might be the first member on which
// handleMergeView() is invoked, which blocks on reception of all view-acks, delaying view installation in
// others. By moving self to the tail, at least the other coords can install the MergeView in parallel
List coords_copy=new ArrayList<>(coords);
if(coords_copy.remove(gms.getAddress()))
coords_copy.add(gms.getAddress());
log.debug("%s: installing merge view %s in %s", gms.getAddress(), combined_merge_data.view.getViewId(), coords_copy);
long start=System.currentTimeMillis();
for(Address coord: coords_copy) {
Message msg=new BytesMessage(coord).setArray(GMS.marshal(view, digest))
.putHeader(gms.getId(), new GMS.GmsHeader(GMS.GmsHeader.INSTALL_MERGE_VIEW).mergeId(merge_id));
gms.getDownProtocol().down(msg);
}
//[JGRP-700] - FLUSH: flushing should span merge; if flush is in stack, wait for acks from subview coordinators
if(gms.flushProtocolInStack) {
try {
gms.merge_ack_collector.waitForAllAcks(gms.view_ack_collection_timeout);
log.trace("%s: received all ACKs (%d) for merge view %s in %d ms",
gms.getAddress(), size, view, (System.currentTimeMillis() - start));
}
catch(TimeoutException e) {
log.warn("%s: failed to collect all ACKs (%d) for merge view %s after %d ms, missing ACKs from %s",
gms.getAddress(), size, view, gms.view_ack_collection_timeout, gms.merge_ack_collector.printMissing());
}
}
}
protected void sendMergeRejectedResponse(Address sender, MergeId merge_id) {
Message msg=new EmptyMessage(sender).setFlag(OOB)
.putHeader(gms.getId(), new GMS.GmsHeader(GMS.GmsHeader.MERGE_RSP).mergeId(merge_id).mergeRejected(true));
gms.getDownProtocol().down(msg);
}
protected void sendMergeCancelledMessage(Collection coords, MergeId merge_id) {
if(coords == null || merge_id == null)
return;
for(Address coord: coords) {
Message msg=new EmptyMessage(coord).putHeader(gms.getId(), new GMS.GmsHeader(GMS.GmsHeader.CANCEL_MERGE).mergeId(merge_id));
gms.getDownProtocol().down(msg);
}
}
/**
* Multicasts a GET_DIGEST_REQ to all members of this sub partition and waits for all responses
* (GET_DIGEST_RSP) or N ms.
*/
protected Digest fetchDigestsFromAllMembersInSubPartition(final View view, MergeId merge_id) {
final List current_mbrs=view.getMembers();
// Optimization: if we're the only member, we don't need to multicast the get-digest message
if(current_mbrs == null || current_mbrs.size() == 1 && current_mbrs.get(0).equals(gms.getAddress()))
return new MutableDigest(view.getMembersRaw())
.set((Digest)gms.getDownProtocol().down(new Event(Event.GET_DIGEST, gms.getAddress())));
Message get_digest_req=new EmptyMessage().setFlag(OOB).setFlag(DONT_LOOPBACK)
.putHeader(gms.getId(), new GMS.GmsHeader(GMS.GmsHeader.GET_DIGEST_REQ).mergeId(merge_id));
long max_wait_time=gms.merge_timeout / 2; // gms.merge_timeout is guaranteed to be > 0, verified in init()
digest_collector.reset(current_mbrs);
long start=System.currentTimeMillis();
gms.getDownProtocol().down(get_digest_req);
// add my own digest first - the get_digest_req needs to be sent first *before* getting our own digest, so
// we have that message in our digest !
Digest digest=(Digest)gms.getDownProtocol().down(new Event(Event.GET_DIGEST, gms.getAddress()));
digest_collector.add(gms.getAddress(), digest);
digest_collector.waitForAllResponses(max_wait_time);
if(log.isTraceEnabled()) {
if(digest_collector.hasAllResponses())
log.trace("%s: fetched all digests for %s in %d ms", gms.getAddress(), current_mbrs, System.currentTimeMillis()-start);
else
log.trace("%s: fetched incomplete digests (after timeout of %d) ms for %s",
gms.getAddress(), max_wait_time, current_mbrs);
}
List valid_rsps=new ArrayList<>(current_mbrs);
valid_rsps.removeAll(digest_collector.getMissing());
Address[] tmp=new Address[valid_rsps.size()];
valid_rsps.toArray(tmp);
MutableDigest retval=new MutableDigest(tmp);
Map responses=new HashMap<>(digest_collector.getResults());
responses.values().forEach(retval::set);
return retval;
}
/**
* Fetches the digests from all members and installs them again. Used only for diagnosis and support; don't
* use this otherwise !
*/
protected void fixDigests() {
Digest digest=fetchDigestsFromAllMembersInSubPartition(gms.view, null);
Message msg=new BytesMessage().putHeader(gms.getId(), new GMS.GmsHeader(GMS.GmsHeader.INSTALL_DIGEST))
.setArray(GMS.marshal(null, digest));
gms.getDownProtocol().down(msg);
}
protected void stop() {
merge_task.stop();
}
protected synchronized void cancelMerge(MergeId id) {
if(setMergeId(id, null)) {
merge_task.stop();
stopMergeKiller();
merge_rsps.reset();
gms.getViewHandler().resume();
gms.getDownProtocol().down(new Event(Event.RESUME_STABLE));
}
}
protected synchronized void forceCancelMerge() {
if(this.merge_id != null)
cancelMerge(this.merge_id);
}
protected synchronized void startMergeKiller() {
if(merge_killer == null || merge_killer.isDone()) {
MergeKiller task=new MergeKiller(this.merge_id);
merge_killer=gms.timer.schedule(task, gms.merge_timeout * 2, TimeUnit.MILLISECONDS, false);
}
}
protected synchronized void stopMergeKiller() {
if(merge_killer != null) {
merge_killer.cancel(false);
merge_killer=null;
}
}
/**
* Starts the merge protocol (only run by the merge leader). Essentially sends a MERGE_REQ to all
* coordinators of all subgroups found. Each coord receives its digest and view and returns it.
* The leader then computes the digest and view for the new group from the return values. Finally, it
* sends this merged view/digest to all subgroup coordinators; each coordinator will install it in their
* subgroup.
*/
protected class MergeTask implements Runnable {
protected Thread thread=null;
/** List of all subpartition coordinators and their members */
protected final ConcurrentMap> coords=Util.createConcurrentMap(8, 0.75f, 8);
protected final Set subviews=new HashSet<>();
/**
* @param views Guaranteed to be non-null and to have >= 2 members, or else this thread would not be started
*/
public synchronized void start(Map views) {
if(thread != null && thread.isAlive()) // the merge thread is already running
return;
this.coords.clear();
this.subviews.clear();
subviews.addAll(views.values());
// now remove all members which don't have us in their view, so RPCs won't block (e.g. FLUSH)
// https://issues.redhat.com/browse/JGRP-1061
sanitizeViews(views);
Map> tmp_coords=determineMergeCoords(views);
this.coords.putAll(tmp_coords);
thread=gms.getThreadFactory().newThread(this, "MergeTask");
thread.setDaemon(true);
thread.start();
}
public synchronized void stop() {
Thread tmp=thread;
if(thread != null && thread.isAlive())
tmp.interrupt();
thread=null;
}
public synchronized boolean isRunning() {
return thread != null && thread.isAlive();
}
public void run() {
// 1. Generate merge_id
final MergeId new_merge_id=MergeId.create(gms.getAddress());
final Collection coordsCopy=new ArrayList<>(coords.keySet());
long start=System.currentTimeMillis();
try {
_run(new_merge_id, coordsCopy); // might remove members from coordsCopy
}
catch(Throwable ex) {
if(ex instanceof Error || ex instanceof RuntimeException)
log.warn(gms.getAddress() + ": merge is cancelled", ex);
else
log.warn("%s: merge is cancelled: %s", gms.getAddress(), ex.getMessage());
sendMergeCancelledMessage(coordsCopy, new_merge_id);
cancelMerge(new_merge_id); // the message above cancels the merge, too, but this is a 2nd line of defense
}
finally {
/* 5. if flush is in stack stop the flush for entire cluster [JGRP-700] - FLUSH: flushing should span merge */
if(gms.flushProtocolInStack)
gms.stopFlush();
thread=null;
}
long diff=System.currentTimeMillis() - start;
log.debug("%s: merge %s took %d ms", gms.getAddress(), new_merge_id, diff);
}
/** Runs the merge protocol as a leader */
protected void _run(MergeId new_merge_id, final Collection coordsCopy) throws Exception {
boolean success=setMergeId(null, new_merge_id);
if(!success) {
log.warn("%s: failed to set my own merge_id (%s) to %s", gms.getAddress(), merge_id, new_merge_id);
return;
}
log.debug("%s: merge task %s started with %d participants", gms.getAddress(), merge_id, coords.keySet().size());
// Fetch the current views and digests from all subgroup coordinators into merge_rsps
success=getMergeDataFromSubgroupCoordinators(coords, new_merge_id, gms.merge_timeout);
List missing=null;
if(!success) {
missing=merge_rsps.getMissing();
log.debug("%s: merge leader %s did not get responses from all %d partition coordinators; " +
"missing responses from %d members, removing them from the merge",
gms.getAddress(), gms.getAddress(), coords.keySet().size(), missing.size());
merge_rsps.remove(missing);
}
// Remove null or rejected merge responses from merge_rsp and coords (so we'll send the new view
// only to members who accepted the merge request)
if(missing != null && !missing.isEmpty()) {
coords.keySet().removeAll(missing);
coordsCopy.removeAll(missing);
}
removeRejectedMergeRequests(coords.keySet());
if(merge_rsps.size() == 0)
throw new Exception("did not get any merge responses from partition coordinators");
if(!coords.containsKey(gms.getAddress())) // another member might have invoked a merge req on us before we got there...
throw new Exception("merge leader rejected merge request");
// Combine all views and digests into 1 View/1 Digest
List merge_data=new ArrayList<>(merge_rsps.getResults().values());
MergeData combined_merge_data=consolidateMergeData(merge_data, new ArrayList<>(subviews));
if(combined_merge_data == null)
throw new Exception("could not consolidate merge");
// send the new view and digest to all coords (including myself). On reception, they will
// install the digest and view in all of their subgroup members
sendMergeView(coords.keySet(), combined_merge_data, new_merge_id);
}
/**
* Sends a MERGE_REQ to all coords and populates a list of MergeData (in merge_rsps). Returns after coords.size()
* response have been received, or timeout msecs have elapsed (whichever is first).
* If a subgroup coordinator rejects the MERGE_REQ (e.g. because of participation in a different merge),
* that member will be removed from coords !
* @param coords A map of coordinatgor addresses and associated membership lists
* @param new_merge_id The new merge id
* @param timeout Max number of msecs to wait for the merge responses from the subgroup coords
*/
protected boolean getMergeDataFromSubgroupCoordinators(Map> coords, MergeId new_merge_id, long timeout) {
boolean gotAllResponses;
long start=System.currentTimeMillis();
merge_rsps.reset(coords.keySet());
log.trace("%s: sending MERGE_REQ to %s", gms.getAddress(), coords.keySet());
for(Map.Entry> entry: coords.entrySet()) {
Address coord=entry.getKey();
Collection mbrs=entry.getValue();
Message msg=new BytesMessage(coord).setFlag(OOB)
.putHeader(gms.getId(), new GMS.GmsHeader(GMS.GmsHeader.MERGE_REQ).mbr(gms.getAddress()).mergeId(new_merge_id))
.setArray(GMS.marshal(mbrs));
gms.getDownProtocol().down(msg);
}
// wait until num_rsps_expected >= num_rsps or timeout elapsed
merge_rsps.waitForAllResponses(timeout);
gotAllResponses=merge_rsps.hasAllResponses();
long time=System.currentTimeMillis() - start;
log.trace("%s: collected %d merge response(s) in %d ms", gms.getAddress(), merge_rsps.numberOfValidResponses(), time);
return gotAllResponses;
}
/** Removed rejected merge requests from merge_rsps and coords. This method has a lock on merge_rsps */
protected void removeRejectedMergeRequests(Collection coords) {
int num_removed=0;
for(Iterator> it=merge_rsps.getResults().entrySet().iterator(); it.hasNext();) {
Map.Entry entry=it.next();
MergeData data=entry.getValue();
if(data.merge_rejected) {
if(data.getSender() != null)
coords.remove(data.getSender());
it.remove();
num_removed++;
}
}
if(num_removed > 0)
log.trace("%s: removed %d rejected merge responses", gms.getAddress(), num_removed);
}
/**
* Merge all MergeData. All MergeData elements should be disjunct (both views and digests). However,
* this method is prepared to resolve duplicate entries (for the same member). The resolution strategy for
* views is to merge only 1 of the duplicate members. Resolution strategy for digests is to take the higher
* seqnos for duplicate digests.
* After merging all members into a Membership and subsequent sorting, the first member of the sorted membership
* will be the new coordinator. This method has a lock on merge_rsps.
* @param merge_rsps A list of MergeData items. Elements with merge_rejected=true were removed before. Is guaranteed
* not to be null and to contain at least 1 member
* @param subviews Contains a list of Views, each View is a subgroup
*/
protected MergeData consolidateMergeData(List merge_rsps, List subviews) {
long logical_time=0; // for new_vid
Collection> sub_mbrships=new ArrayList<>();
Set digest_membership=new HashSet<>(); // members as seen by the digests
for(MergeData tmp_data: merge_rsps) {
View tmp_view=tmp_data.getView();
if(tmp_view != null) {
ViewId tmp_vid=tmp_view.getViewId();
if(tmp_vid != null)
logical_time=Math.max(logical_time, tmp_vid.getId()); // compute the new view id (max of all vids +1)
// merge all membership lists into one (prevent duplicates)
sub_mbrships.add(new ArrayList<>(tmp_view.getMembers()));
}
Digest digest=tmp_data.getDigest();
if(digest != null)
for(Digest.Entry entry: digest)
digest_membership.add(entry.getMember());
}
// remove all members from the new view for which we didn't get a digest, e.g. new view={A,B,C,D,E,F},
// digest={A,C,D,F} --> new view={A,C,D,F}, digest={A,C,D,F}
if(!digest_membership.isEmpty())
for(Collection coll: sub_mbrships)
coll.retainAll(digest_membership);
List merged_mbrs=gms.computeNewMembership(sub_mbrships);
// remove all members from the (future) MergeView that are not in the digest
// Usually not needed, but a second line of defense in case the membership change policy above
// computed the new view incorrectly, e.g. not removing dupes
Set all_members=new HashSet<>();
sub_mbrships.forEach(all_members::addAll);
merged_mbrs.retainAll(all_members);
// the new coordinator is the first member of the consolidated & sorted membership list
Address new_coord=merged_mbrs.isEmpty()? null : merged_mbrs.get(0);
if(new_coord == null)
return null;
// Remove views from subviews whose creator are not in the new membership
for(Iterator it=subviews.iterator(); it.hasNext();) {
View v=it.next();
logical_time=Math.max(logical_time, v.getViewId().getId());
Address creator=v.getCreator();
if(creator != null && !merged_mbrs.contains(creator))
it.remove();
}
// determine the new view; logical_time should be the highest view ID seen up to now plus 1
MergeView new_view=new MergeView(new_coord, logical_time + 1, merged_mbrs, subviews);
// determine the new digest
MutableDigest new_digest=consolidateDigests(new_view, merge_rsps);
if(new_digest == null || !new_digest.allSet())
return null;
log.trace("%s: consolidated view=%s\nconsolidated digest=%s", gms.getAddress(), new_view, new_digest);
return new MergeData(gms.getAddress(), new_view, new_digest);
}
/**
* Merge all digests into one. For each sender, the new value is max(highest_delivered),
* max(highest_received). This method has a lock on merge_rsps
*/
protected MutableDigest consolidateDigests(final View new_view, final List merge_rsps) {
MutableDigest retval=new MutableDigest(new_view.getMembersRaw());
for(MergeData data: merge_rsps) {
Digest tmp_digest=data.getDigest();
if(tmp_digest != null)
retval.merge(tmp_digest);
}
return retval;
}
}
protected class MergeKiller implements Runnable {
protected final MergeId my_merge_id;
MergeKiller(MergeId my_merge_id) {
this.my_merge_id=my_merge_id;
}
public void run() {
cancelMerge(my_merge_id);
}
public String toString() {
return Merger.class.getSimpleName() + ": " + getClass().getSimpleName();
}
}
}