/**
* Sends the new view and digest to all subgroup coordinors in coords. Each coord will in turn
*
* <ol>
* <li>broadcast the new view and digest to all the members of its subgroup (MergeView)
* <li>on reception of the view, if it is a MergeView, each member will set the digest and
* install the new view
* </ol>
*/
private void sendMergeView(
Collection<Address> coords, MergeData combined_merge_data, MergeId merge_id) {
if (coords == null || combined_merge_data == null) return;
View view = combined_merge_data.view;
Digest digest = combined_merge_data.digest;
if (view == null || digest == null) {
if (log.isErrorEnabled())
log.error("view or digest is null, cannot send consolidated merge view/digest");
return;
}
if (log.isDebugEnabled())
log.debug(
gms.local_addr + ": sending merge view " + view.getVid() + " to coordinators " + coords);
gms.merge_ack_collector.reset(coords);
int size = gms.merge_ack_collector.size();
long timeout = gms.view_ack_collection_timeout;
long start = System.currentTimeMillis();
for (Address coord : coords) {
Message msg = new Message(coord, null, null);
GMS.GmsHeader hdr = new GMS.GmsHeader(GMS.GmsHeader.INSTALL_MERGE_VIEW);
hdr.view = view;
hdr.my_digest = digest;
hdr.merge_id = merge_id;
msg.putHeader(gms.getId(), hdr);
gms.getDownProtocol().down(new Event(Event.MSG, msg));
}
// [JGRP-700] - FLUSH: flushing should span merge
// if flush is in stack wait for acks from separated island coordinators
if (gms.flushProtocolInStack) {
try {
gms.merge_ack_collector.waitForAllAcks(timeout);
long stop = System.currentTimeMillis();
if (log.isTraceEnabled())
log.trace(
"received all ACKs ("
+ size
+ ") for merge view "
+ view
+ " in "
+ (stop - start)
+ "ms");
} catch (TimeoutException e) {
log.warn(
gms.local_addr
+ ": failed to collect all ACKs for merge ("
+ size
+ ") for view "
+ view
+ " after "
+ timeout
+ "ms, missing ACKs from "
+ gms.merge_ack_collector.printMissing());
}
}
}
/**
* 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).
*
* <p>If a subgroup coordinator rejects the MERGE_REQ (e.g. because of participation in a
* different merge), <em>that member will be removed from coords !</em>
*
* @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
*/
private boolean getMergeDataFromSubgroupCoordinators(
Map<Address, Collection<Address>> coords, MergeId new_merge_id, long timeout) {
boolean gotAllResponses;
long start = System.currentTimeMillis();
merge_rsps.reset(coords.keySet());
if (log.isDebugEnabled())
log.debug(gms.local_addr + ": sending MERGE_REQ to " + coords.keySet());
for (Map.Entry<Address, Collection<Address>> entry : coords.entrySet()) {
Address coord = entry.getKey();
Collection<Address> mbrs = entry.getValue();
Message msg = new Message(coord, null, null);
msg.setFlag(Message.OOB);
GMS.GmsHeader hdr = new GMS.GmsHeader(GMS.GmsHeader.MERGE_REQ, mbrs);
hdr.mbr = gms.local_addr;
hdr.merge_id = new_merge_id;
msg.putHeader(gms.getId(), hdr);
gms.getDownProtocol().down(new Event(Event.MSG, msg));
}
// wait until num_rsps_expected >= num_rsps or timeout elapsed
merge_rsps.waitForAllResponses(timeout);
gotAllResponses = merge_rsps.hasAllResponses();
long stop = System.currentTimeMillis();
if (log.isDebugEnabled())
log.debug(
gms.local_addr
+ ": collected "
+ merge_rsps.size()
+ " merge response(s) in "
+ (stop - start)
+ " ms");
return gotAllResponses;
}
/**
* Fetches the digests from all members and installs them again. Used only for diagnosis and
* support; don't use this otherwise !
*/
void fixDigests() {
Digest digest = fetchDigestsFromAllMembersInSubPartition(gms.view.getMembers());
Message msg = new Message();
GMS.GmsHeader hdr = new GMS.GmsHeader(GMS.GmsHeader.INSTALL_DIGEST);
hdr.my_digest = digest;
msg.putHeader(gms.getId(), hdr);
gms.getDownProtocol().down(new Event(Event.MSG, msg));
}
protected void sendMergeRejectedResponse(Address sender, MergeId merge_id) {
Message msg = new Message(sender).setFlag(Message.Flag.OOB, Message.Flag.INTERNAL);
GMS.GmsHeader hdr = new GMS.GmsHeader(GMS.GmsHeader.MERGE_RSP);
hdr.merge_rejected = true;
hdr.merge_id = merge_id;
msg.putHeader(gms.getId(), hdr);
gms.getDownProtocol().down(new Event(Event.MSG, msg));
}
protected void sendMergeRejectedResponse(Address sender, MergeId merge_id) {
Message msg = new Message(sender, null, null);
msg.setFlag(Message.OOB);
GMS.GmsHeader hdr = new GMS.GmsHeader(GMS.GmsHeader.MERGE_RSP);
hdr.merge_rejected = true;
hdr.merge_id = merge_id;
msg.putHeader(gms.getId(), hdr);
if (log.isDebugEnabled()) log.debug("merge response=" + hdr);
gms.getDownProtocol().down(new Event(Event.MSG, msg));
}
/** Send back a response containing view and digest to sender */
private void sendMergeResponse(Address sender, View view, Digest digest, MergeId merge_id) {
Message msg = new Message(sender).setFlag(Message.Flag.OOB, Message.Flag.INTERNAL);
GMS.GmsHeader hdr = new GMS.GmsHeader(GMS.GmsHeader.MERGE_RSP);
hdr.merge_id = merge_id;
hdr.view = view;
hdr.my_digest = digest;
msg.putHeader(gms.getId(), hdr);
if (log.isTraceEnabled()) log.trace(gms.local_addr + ": sending merge response=" + hdr);
gms.getDownProtocol().down(new Event(Event.MSG, msg));
}
private void sendMergeCancelledMessage(Collection<Address> coords, MergeId merge_id) {
if (coords == null || merge_id == null) return;
for (Address coord : coords) {
Message msg = new Message(coord);
// msg.setFlag(Message.Flag.OOB);
GMS.GmsHeader hdr = new GMS.GmsHeader(GMS.GmsHeader.CANCEL_MERGE);
hdr.merge_id = merge_id;
msg.putHeader(gms.getId(), hdr);
gms.getDownProtocol().down(new Event(Event.MSG, msg));
}
}
private void sendMergeCancelledMessage(Collection<Address> coords, MergeId merge_id) {
if (coords == null || merge_id == null) return;
for (Address coord : coords) {
Message msg = new Message(coord, null, null);
// msg.setFlag(Message.OOB);
GMS.GmsHeader hdr = new GMS.GmsHeader(GMS.GmsHeader.CANCEL_MERGE);
hdr.merge_id = merge_id;
msg.putHeader(gms.getId(), hdr);
if (log.isDebugEnabled()) log.debug(gms.local_addr + ": sending cancel merge to " + coord);
gms.getDownProtocol().down(new Event(Event.MSG, msg));
}
}
/**
* Multicasts a GET_DIGEST_REQ to all current members and waits for all responses (GET_DIGEST_RSP)
* or N ms.
*
* @return
*/
private Digest fetchDigestsFromAllMembersInSubPartition(
List<Address> current_mbrs, MergeId merge_id) {
// 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.local_addr))
return (Digest) gms.getDownProtocol().down(new Event(Event.GET_DIGEST, gms.local_addr));
GMS.GmsHeader hdr = new GMS.GmsHeader(GMS.GmsHeader.GET_DIGEST_REQ);
hdr.merge_id = merge_id;
Message get_digest_req =
new Message().setFlag(Message.Flag.OOB, Message.Flag.INTERNAL).putHeader(gms.getId(), hdr);
long max_wait_time =
gms.merge_timeout / 2; // gms.merge_timeout is guaranteed to be > 0, verified in init()
digest_collector.reset(current_mbrs);
gms.getDownProtocol().down(new Event(Event.MSG, 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.local_addr));
digest_collector.add(gms.local_addr, digest);
digest_collector.waitForAllResponses(max_wait_time);
if (log.isTraceEnabled()) {
if (digest_collector.hasAllResponses())
log.trace(gms.local_addr + ": fetched all digests for " + current_mbrs);
else
log.trace(
gms.local_addr
+ ": fetched incomplete digests (after timeout of "
+ max_wait_time
+ ") ms for "
+ current_mbrs);
}
Map<Address, Digest> responses = new HashMap<Address, Digest>(digest_collector.getResults());
MutableDigest retval = new MutableDigest(responses.size());
for (Digest dig : responses.values()) {
if (dig != null) retval.add(dig);
}
return retval;
}