echelon/i2p.i2p
1
0
Fork 0
forked from I2P_Developers/i2p.i2p
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

- Fix first hop expiration for Build Messages, was way too long

Browse Source 
- Randomize Build Message expiration to make it harder to guess hop position
      - Save expired tunnel build configs for a while, so that we will still use the tunnel
        and update peer stats if the reply comes in late
      - Don't update our own profile for Tunnel Build Replies
This commit is contained in:
zzz
2010-02-01 14:17:11 +00:00
parent 839986db22
commit 25e51a945c
3 changed files with 70 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

64
router/java/src/net/i2p/router/tunnel/pool/BuildExecutor.java
View File

@ -34,10 +34,14 @@ class BuildExecutor implements Runnable {
private final Object _currentlyBuilding;
/** indexed by ptcc.getReplyMessageId() */
private final ConcurrentHashMap<Long, PooledTunnelCreatorConfig> _currentlyBuildingMap;
/** indexed by ptcc.getReplyMessageId() */
private final ConcurrentHashMap<Long, PooledTunnelCreatorConfig> _recentlyBuildingMap;
private boolean _isRunning;
private BuildHandler _handler;
private boolean _repoll;
private static final int MAX_CONCURRENT_BUILDS = 10;
/** accept replies up to a minute after we gave up on them */
private static final long GRACE_PERIOD = 60*1000;
public BuildExecutor(RouterContext ctx, TunnelPoolManager mgr) {
_context = ctx;
@ -45,6 +49,7 @@ class BuildExecutor implements Runnable {
_manager = mgr;
_currentlyBuilding = new Object();
_currentlyBuildingMap = new ConcurrentHashMap(MAX_CONCURRENT_BUILDS);
_recentlyBuildingMap = new ConcurrentHashMap(4 * MAX_CONCURRENT_BUILDS);
_context.statManager().createRateStat("tunnel.concurrentBuilds", "How many builds are going at once", "Tunnels", new long[] { 60*1000, 5*60*1000, 60*60*1000 });
_context.statManager().createRateStat("tunnel.concurrentBuildsLagged", "How many builds are going at once when we reject further builds, due to job lag (period is lag)", "Tunnels", new long[] { 60*1000, 5*60*1000, 60*60*1000 });
_context.statManager().createRateStat("tunnel.buildExploratoryExpire", "How often an exploratory tunnel times out during creation", "Tunnels", new long[] { 10*60*1000, 60*60*1000 });
@ -57,6 +62,7 @@ class BuildExecutor implements Runnable {
_context.statManager().createRateStat("tunnel.buildRequestZeroHopTime", "How long it takes to build a zero hop tunnel", "Tunnels", new long[] { 60*1000, 10*60*1000 });
_context.statManager().createRateStat("tunnel.pendingRemaining", "How many inbound requests are pending after a pass (period is how long the pass takes)?", "Tunnels", new long[] { 60*1000, 10*60*1000 });
_context.statManager().createRateStat("tunnel.buildFailFirstHop", "How often we fail to build a OB tunnel because we can't contact the first hop", "Tunnels", new long[] { 60*1000, 10*60*1000 });
_context.statManager().createRateStat("tunnel.buildReplySlow", "Build reply late, but not too late", "Tunnels", new long[] { 10*60*1000 });
// Get stat manager, get recognized bandwidth tiers
StatManager statMgr = _context.statManager();
@ -84,22 +90,33 @@ class BuildExecutor implements Runnable {
if (allowed > MAX_CONCURRENT_BUILDS) allowed = MAX_CONCURRENT_BUILDS; // Never go beyond 10, that is uncharted territory (old limit was 5)
allowed = _context.getProperty("router.tunnelConcurrentBuilds", allowed);
// expire any REALLY old requests
long expireBefore = _context.clock().now() + 10*60*1000 - BuildRequestor.REQUEST_TIMEOUT - GRACE_PERIOD;
for (Iterator<PooledTunnelCreatorConfig> iter = _recentlyBuildingMap.values().iterator(); iter.hasNext(); ) {
PooledTunnelCreatorConfig cfg = iter.next();
if (cfg.getExpiration() <= expireBefore) {
iter.remove();
}
}
// expire any old requests
List<PooledTunnelCreatorConfig> expired = null;
int concurrent = 0;
// Todo: Make expiration variable
long expireBefore = _context.clock().now() + 10*60*1000 - BuildRequestor.REQUEST_TIMEOUT;
// expire any old requests
for (Iterator<PooledTunnelCreatorConfig> iter = _currentlyBuildingMap.values().iterator(); iter.hasNext(); ) {
PooledTunnelCreatorConfig cfg = iter.next();
if (cfg.getExpiration() <= expireBefore) {
iter.remove();
if (expired == null)
expired = new ArrayList();
expired.add(cfg);
}
expireBefore = _context.clock().now() + 10*60*1000 - BuildRequestor.REQUEST_TIMEOUT;
for (Iterator<PooledTunnelCreatorConfig> iter = _currentlyBuildingMap.values().iterator(); iter.hasNext(); ) {
PooledTunnelCreatorConfig cfg = iter.next();
if (cfg.getExpiration() <= expireBefore) {
// save them for another minute
_recentlyBuildingMap.putIfAbsent(Long.valueOf(cfg.getReplyMessageId()), cfg);
iter.remove();
if (expired == null)
expired = new ArrayList();
expired.add(cfg);
}
concurrent = _currentlyBuildingMap.size();
allowed -= concurrent;
}
concurrent = _currentlyBuildingMap.size();
allowed -= concurrent;
if (expired != null) {
for (int i = 0; i < expired.size(); i++) {
@ -114,7 +131,7 @@ class BuildExecutor implements Runnable {
// Look up peer
Hash peer = cfg.getPeer(iPeer);
// Avoid recording ourselves
if (peer.toBase64().equals(_context.routerHash().toBase64()))
if (peer.equals(_context.routerHash()))
continue;
// Look up routerInfo
RouterInfo ri = _context.netDb().lookupRouterInfoLocally(peer);
@ -496,11 +513,30 @@ class BuildExecutor implements Runnable {
}
/**
* This returns the PTCC up to a minute after it 'expired', thus allowing us to
* still use a tunnel if it was accepted, and to update peer stats.
* This means that manager.buildComplete() could be called more than once, and
* a build can be failed or successful after it was timed out,
* which will affect the stats and profiles.
* But that's ok. A peer that rejects slowly gets penalized twice, for example.
*
* @return ptcc or null
*/
PooledTunnelCreatorConfig removeFromBuilding(long id) {
//_log.error("Removing ID: " + id + "; size was: " + _currentlyBuildingMap.size());
return _currentlyBuildingMap.remove(Long.valueOf(id));
Long key = Long.valueOf(id);
PooledTunnelCreatorConfig rv = _currentlyBuildingMap.remove(key);
if (rv != null)
return rv;
rv = _recentlyBuildingMap.remove(key);
if (rv != null) {
long requestedOn = rv.getExpiration() - 10*60*1000;
long rtt = _context.clock().now() - requestedOn;
_context.statManager().addRateData("tunnel.buildReplySlow", rtt, 0);
if (_log.shouldLog(Log.WARN))
_log.warn("Got reply late (rtt = " + rtt + ") for: " + rv);
}
return rv;
}
public int getInboundBuildQueueSize() { return _handler.getInboundBuildQueueSize(); }

11
router/java/src/net/i2p/router/tunnel/pool/BuildHandler.java
View File

@ -253,6 +253,12 @@ class BuildHandler {
// For each peer in the tunnel
for (int i = 0; i < cfg.getLength(); i++) {
Hash peer = cfg.getPeer(i);
// If this tunnel member is us, skip this record, don't update profile or stats
// for ourselves, we always agree
// Why must we save a slot for ourselves anyway?
if (peer.equals(_context.routerHash()))
continue;
int record = order.indexOf(Integer.valueOf(i));
if (record < 0) {
_log.error("Bad status index " + i);
@ -260,9 +266,9 @@ class BuildHandler {
_exec.buildComplete(cfg, cfg.getTunnelPool());
return;
}
int howBad = statuses[record];
// If this tunnel member isn't ourselves
if (!peer.toBase64().equals(_context.routerHash().toBase64())) {
// Look up routerInfo
RouterInfo ri = _context.netDb().lookupRouterInfoLocally(peer);
// Default and detect bandwidth tier
@ -277,7 +283,6 @@ class BuildHandler {
}
if (_log.shouldLog(Log.INFO))
_log.info(msg.getUniqueId() + ": Peer " + peer.toBase64() + " replied with status " + howBad);
}
if (howBad == 0) {
// w3wt

14
router/java/src/net/i2p/router/tunnel/pool/BuildRequestor.java
View File

@ -42,7 +42,13 @@ class BuildRequestor {
*
*/
static final int REQUEST_TIMEOUT = 13*1000;
/** make this shorter than REQUEST_TIMEOUT */
private static final int FIRST_HOP_TIMEOUT = 10*1000;
/** some randomization is added on to this */
private static final int BUILD_MSG_TIMEOUT = 60*1000;
private static boolean usePairedTunnels(RouterContext ctx) {
String val = ctx.getProperty("router.usePairedTunnels");
if ( (val == null) || (Boolean.valueOf(val).booleanValue()) )
@ -144,8 +150,12 @@ class BuildRequestor {
+ " with msgId=" + msg.getUniqueId());
// send it directly to the first hop
OutNetMessage outMsg = new OutNetMessage(ctx);
// Todo: add some fuzz to the expiration to make it harder to guess how many hops?
outMsg.setExpiration(msg.getMessageExpiration());
// Add some fuzz to the TBM expiration to make it harder to guess how many hops
// or placement in the tunnel
msg.setMessageExpiration(ctx.clock().now() + BUILD_MSG_TIMEOUT + ctx.random().nextLong(20*1000));
// We set the OutNetMessage expiration much shorter, so that the
// TunnelBuildFirstHopFailJob fires before the 13s build expiration.
outMsg.setExpiration(ctx.clock().now() + FIRST_HOP_TIMEOUT);
outMsg.setMessage(msg);
outMsg.setPriority(PRIORITY);
RouterInfo peer = ctx.netDb().lookupRouterInfoLocally(cfg.getPeer(1));