You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
yacy_search_server/htroot/yacy/transferRWI.java

288 lines
14 KiB

// transferRWI.java
// -----------------------
// part of the AnomicHTTPD caching proxy
// (C) by Michael Peter Christen; mc@yacy.net
// first published on http://www.anomic.de
// Frankfurt, Germany, 2004, 2005
//
// $LastChangedDate$
// $LastChangedRevision$
// $LastChangedBy$
//
// This program is free software; you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation; either version 2 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License
// along with this program; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
// You must compile this file with
// javac -classpath .:../classes transferRWI.java
*) Asynchronous queuing of crawl job URLs (stackCrawl) various checks like the blacklist check or the robots.txt disallow check are now done by a separate thread to unburden the indexer thread(s) TODO: maybe we have to introduce a threadpool here if it turn out that this single thread is a bottleneck because of the time consuming robots.txt downloads *) improved index transfer The index selection and transmission is done in parallel now to improve index transfer performance. TODO: maybe we could speed up performance by unsing multiple transmission threads in parallel instead of only a single one. *) gzip encoded post requests it is now configureable if a gzip encoded post request should be send on intex transfer/distribution *) storage Peer (very experimentell and not optimized yet) Now it's possible to send the result of the yacy indexer thread to a remote peer istead of storing the indexed words locally. This could be done by setting the property "storagePeerHash" in the yacy config file - Please note that if the index transfer fails, the index ist stored locally. - TODO: currently this index transfer is done by the indexer thread. To seedup the indexer a) this transmission should be done in parallel and b) multiple chunks should be bundled and transfered together *) general performance improvements - better memory cleanup after http request processing has finished - replacing some string concatenations with stringBuffers - replacing BufferedInputStreams with serverByteBuffer - replacing vectors with arraylists wherever possible - replacing hashtables with hashmaps wherever possible This was done because function calls to verctor or hashtable functions take 3 time longer than calls to functions of arraylists or hashmaps. TODO: we should take a look on the class serverObject which is inherited from hashmap Do we realy need a synchronization for this class? TODO: replace arraylists with linkedLists if random access to the list elements is not needed *) Robots Parser supports if-modified-since downloads now If the downloaded robots.txt file is older than 7 days the robots parser tries to download the robots.txt with the if-modified-since header to avoid unnecessary downloads if the file was not changed. Additionally the ETag header is used to detect changes. *) Crawler: better handling of unsupported mimeTypes + FileExtension *) Bugfix: plasmaWordIndexEntity was not closed correctly in - query.java - plasmaswitchboard.java *) function minimizeUrlDB added to yacy.java this function tests the current urlHashDB for unused urls ATTENTION: please don't use this function at the moment because it causes the wordIndexDB to flush all words into the word directory! git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@853 6c8d7289-2bf4-0310-a012-ef5d649a1542
19 years ago
import java.io.IOException;
import java.util.ArrayList;
import java.util.HashSet;
import java.util.Iterator;
import java.util.Set;
import net.yacy.cora.document.encoding.ASCII;
import net.yacy.cora.document.encoding.UTF8;
import net.yacy.cora.document.feed.RSSMessage;
12 years ago
import net.yacy.cora.federate.yacy.Distribution;
import net.yacy.cora.protocol.HeaderFramework;
import net.yacy.cora.protocol.RequestHeader;
import net.yacy.cora.storage.HandleSet;
import net.yacy.cora.util.ConcurrentLog;
import net.yacy.cora.util.Memory;
import net.yacy.cora.util.SpaceExceededException;
import net.yacy.kelondro.data.word.Word;
import net.yacy.kelondro.data.word.WordReferenceRow;
import net.yacy.kelondro.index.RowHandleSet;
import net.yacy.kelondro.util.FileUtils;
import net.yacy.kelondro.util.MemoryControl;
import net.yacy.peers.EventChannel;
import net.yacy.peers.Network;
import net.yacy.peers.Protocol;
import net.yacy.peers.Seed;
import net.yacy.repository.Blacklist.BlacklistType;
import net.yacy.search.Switchboard;
import net.yacy.search.SwitchboardConstants;
12 years ago
import net.yacy.server.serverObjects;
import net.yacy.server.serverSwitch;
*) Asynchronous queuing of crawl job URLs (stackCrawl) various checks like the blacklist check or the robots.txt disallow check are now done by a separate thread to unburden the indexer thread(s) TODO: maybe we have to introduce a threadpool here if it turn out that this single thread is a bottleneck because of the time consuming robots.txt downloads *) improved index transfer The index selection and transmission is done in parallel now to improve index transfer performance. TODO: maybe we could speed up performance by unsing multiple transmission threads in parallel instead of only a single one. *) gzip encoded post requests it is now configureable if a gzip encoded post request should be send on intex transfer/distribution *) storage Peer (very experimentell and not optimized yet) Now it's possible to send the result of the yacy indexer thread to a remote peer istead of storing the indexed words locally. This could be done by setting the property "storagePeerHash" in the yacy config file - Please note that if the index transfer fails, the index ist stored locally. - TODO: currently this index transfer is done by the indexer thread. To seedup the indexer a) this transmission should be done in parallel and b) multiple chunks should be bundled and transfered together *) general performance improvements - better memory cleanup after http request processing has finished - replacing some string concatenations with stringBuffers - replacing BufferedInputStreams with serverByteBuffer - replacing vectors with arraylists wherever possible - replacing hashtables with hashmaps wherever possible This was done because function calls to verctor or hashtable functions take 3 time longer than calls to functions of arraylists or hashmaps. TODO: we should take a look on the class serverObject which is inherited from hashmap Do we realy need a synchronization for this class? TODO: replace arraylists with linkedLists if random access to the list elements is not needed *) Robots Parser supports if-modified-since downloads now If the downloaded robots.txt file is older than 7 days the robots parser tries to download the robots.txt with the if-modified-since header to avoid unnecessary downloads if the file was not changed. Additionally the ETag header is used to detect changes. *) Crawler: better handling of unsupported mimeTypes + FileExtension *) Bugfix: plasmaWordIndexEntity was not closed correctly in - query.java - plasmaswitchboard.java *) function minimizeUrlDB added to yacy.java this function tests the current urlHashDB for unused urls ATTENTION: please don't use this function at the moment because it causes the wordIndexDB to flush all words into the word directory! git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@853 6c8d7289-2bf4-0310-a012-ef5d649a1542
19 years ago
public final class transferRWI {
public static serverObjects respond(final RequestHeader header, final serverObjects post, final serverSwitch env) {
// return variable that accumulates replacements
final Switchboard sb = (Switchboard) env;
// remember the peer contact for peer statistics
final String clientip = header.getRemoteAddr();
final String userAgent = header.get(HeaderFramework.USER_AGENT, "<unknown>");
if (clientip != null) sb.peers.peerActions.setUserAgent(clientip, userAgent);
final serverObjects prop = new serverObjects();
final String contentType = header.getContentType();
prop.put("unknownURL", "");
prop.put("pause", 60000);
String result = "";
if ((post == null) || (env == null)) {
result = "post or env is null!";
logWarning(contentType, result);
prop.put("result", result);
return prop;
}
if (!Protocol.authentifyRequest(post, env)) {
result = "not authentified";
prop.put("result", result);
return prop;
}
if (!post.containsKey("wordc")) {
result = "missing wordc";
prop.put("result", result);
return prop;
}
if (!post.containsKey("entryc")) {
result = "missing entryc";
prop.put("result", result);
return prop;
}
if (!post.containsKey("indexes")) {
result = "missing indexes";
prop.put("result", result);
return prop;
}
// load tests
if (Memory.load() > 2.0 || MemoryControl.shortStatus()) {
// check also Protocol.metadataRetrievalRunning.get() > 0 ?
result = "too high load"; // don't tell too much details
prop.put("result", result);
return prop;
}
// request values
final String iam = post.get("iam", ""); // seed hash of requester
final String youare = post.get("youare", ""); // seed hash of the target peer, needed for network stability
// final String key = (String) post.get("key", ""); // transmission key
final int wordc = post.getInt("wordc", 0); // number of different words
final int entryc = post.getInt("entryc", 0); // number of entries in indexes
byte[] indexes = post.get("indexes", "").getBytes(); // the indexes, as list of word entries
boolean granted = sb.getConfigBool(SwitchboardConstants.INDEX_RECEIVE_ALLOW, false);
final boolean blockBlacklist = sb.getConfigBool(SwitchboardConstants.INDEX_RECEIVE_BLOCK_BLACKLIST, false);
final long cachelimit = sb.getConfigLong(SwitchboardConstants.WORDCACHE_MAX_COUNT, 100000);
final Seed otherPeer = sb.peers.get(iam);
final String otherPeerName = iam + ":" + ((otherPeer == null) ? "NULL" : (otherPeer.getName() + "/" + otherPeer.getVersion()));
// response values
int pause = 0;
result = "ok";
final StringBuilder unknownURLs = new StringBuilder(6000);
double load = Memory.load();
float maxload = sb.getConfigFloat(SwitchboardConstants.INDEX_DIST_LOADPREREQ, 2.0f);
if (load > maxload) {
// too high local load. this is bad but we must reject this to protect ourself!
sb.getLog().info("Rejecting RWIs from peer " + otherPeerName + ", system has too high load = " + load + ", maxload = " + maxload);
result = "not_granted";
pause = (int) (load * 20000);
} else if ((youare == null) || (!youare.equals(sb.peers.mySeed().hash))) {
sb.getLog().info("Rejecting RWIs from peer " + otherPeerName + ". Wrong target. Wanted peer=" + youare + ", iam=" + sb.peers.mySeed().hash);
result = "wrong_target";
pause = 0;
} else if (otherPeer == null) {
// we dont want to receive indexes
sb.getLog().info("Rejecting RWIs from peer " + otherPeerName + ". Not granted. Other Peer is unknown");
result = "not_granted";
pause = 60000;
} else if (!granted) {
// we dont want to receive indexes
sb.getLog().info("Rejecting RWIs from peer " + otherPeerName + ". Granted is false");
result = "not_granted";
pause = 60000;
} else if (sb.isRobinsonMode()) {
// we dont want to receive indexes
sb.getLog().info("Rejecting RWIs from peer " + otherPeerName + ". Not granted. This peer is in robinson mode");
result = "not_granted";
pause = 60000;
} else if (sb.index.RWIBufferCount() > cachelimit) {
// we are too busy to receive indexes
sb.getLog().info("Rejecting RWIs from peer " + otherPeerName + ". We are too busy (buffersize=" + sb.index.RWIBufferCount() + ").");
granted = false; // don't accept more words if there are too many words to flush
result = "busy";
pause = 60000;
} else if (otherPeer.getVersion() < 0.75005845 && otherPeer.getVersion() >= 0.75005821) {
// version that sends [B@... hashes
sb.getLog().info("Rejecting RWIs from peer " + otherPeerName + ". Bad version.");
result = "not_granted";
pause = 1800000;
} else {
// we want and can receive indexes
// log value status (currently added to find outOfMemory error
if (sb.getLog().isFine()) sb.getLog().fine("Processing " + indexes.length + " bytes / " + wordc + " words / " + entryc + " entries from " + otherPeerName);
final long startProcess = System.currentTimeMillis();
// decode request
//System.out.println("STRINGS " + UTF8.String(indexes));
final Iterator<String> it = FileUtils.strings(indexes);
// free memory
indexes = null;
// now parse the Strings in the value-vector and write index entries
String estring;
int p;
String wordHash;
byte[] urlHash;
WordReferenceRow iEntry;
final HandleSet unknownURL = new RowHandleSet(Word.commonHashLength, Word.commonHashOrder, 0);
final ArrayList<String> wordhashes = new ArrayList<String>();
int received = 0;
int blocked = 0;
int count = 0;
Set<String> testids = new HashSet<String>();
while (it.hasNext()) {
estring = it.next();
count++;
if (count > 1000) break; // protection against flooding
// check if RWI entry is well-formed
p = estring.indexOf('{',0);
if (p < 0 || estring.indexOf("x=",0) < 0 || !(estring.indexOf("[B@",0) < 0)) {
blocked++;
continue;
}
wordHash = estring.substring(0, p);
wordhashes.add(wordHash);
iEntry = new WordReferenceRow(estring.substring(p));
urlHash = iEntry.urlhash();
// block blacklisted entries
if ((blockBlacklist) && (Switchboard.urlBlacklist.hashInBlacklistedCache(BlacklistType.DHT, urlHash))) {
Network.log.fine("transferRWI: blocked blacklisted URLHash '" + ASCII.String(urlHash) + "' from peer " + otherPeerName);
blocked++;
continue;
}
// check if the entry is in our network domain
final String urlRejectReason = sb.crawlStacker.urlInAcceptedDomainHash(urlHash);
if (urlRejectReason != null) {
Network.log.warn("transferRWI: blocked URL hash '" + ASCII.String(urlHash) + "' (" + urlRejectReason + ") from peer " + otherPeerName + "; peer is suspected to be a spam-peer (or something is wrong)");
//if (yacyCore.log.isFine()) yacyCore.log.logFine("transferRWI: blocked URL hash '" + urlHash + "' (" + urlRejectReason + ") from peer " + otherPeerName);
blocked++;
continue;
}
// learn entry
try {
sb.index.storeRWI(ASCII.getBytes(wordHash), iEntry);
} catch (final Exception e) {
ConcurrentLog.logException(e);
}
// check if we need to ask for the corresponding URL
testids.add(ASCII.String(urlHash));
received++;
}
for (String id: testids) {
try {
try {
if (sb.index.fulltext().getLoadTime(id) < 0) {
unknownURL.put(ASCII.getBytes(id));
}
} catch (IOException e) {
ConcurrentLog.logException(e);
unknownURL.put(ASCII.getBytes(id));
}
} catch (final SpaceExceededException e) {
sb.getLog().warn("transferRWI: DB-Error while trying to determine if URL with hash '" + id + "' is known.", e);
}
}
sb.peers.mySeed().incRI(received);
// finally compose the unknownURL hash list
final Iterator<byte[]> bit = unknownURL.iterator();
unknownURLs.ensureCapacity(unknownURL.size() * 25);
while (bit.hasNext()) {
unknownURLs.append(UTF8.String(bit.next())).append(',');
}
if (unknownURLs.length() > 0) { unknownURLs.setLength(unknownURLs.length() - 1); }
if (wordhashes.isEmpty() || received == 0) {
sb.getLog().info("Received 0 RWIs from " + otherPeerName + ", processed in " + (System.currentTimeMillis() - startProcess) + " milliseconds, requesting " + unknownURL.size() + " URLs, blocked " + blocked + " RWIs");
} else {
final String firstHash = wordhashes.get(0);
final String lastHash = wordhashes.get(wordhashes.size() - 1);
final long avdist = (Distribution.horizontalDHTDistance(firstHash.getBytes(), ASCII.getBytes(sb.peers.mySeed().hash)) + Distribution.horizontalDHTDistance(lastHash.getBytes(), ASCII.getBytes(sb.peers.mySeed().hash))) / 2;
sb.getLog().info("Received " + received + " RWIs, " + wordc + " Words [" + firstHash + " .. " + lastHash + "], processed in " + (System.currentTimeMillis() - startProcess) + " milliseconds, " + avdist + ", blocked " + blocked + ", requesting " + unknownURL.size() + "/" + received+ " URLs from " + otherPeerName);
EventChannel.channels(EventChannel.DHTRECEIVE).addMessage(new RSSMessage("Received " + received + " RWIs, " + wordc + " Words [" + firstHash + " .. " + lastHash + "], processed in " + (System.currentTimeMillis() - startProcess) + " milliseconds, " + avdist + ", blocked " + blocked + ", requesting " + unknownURL.size() + "/" + received + " URLs from " + otherPeerName, "", otherPeer.hash));
}
result = "ok";
pause = (int) (sb.index.RWIBufferCount() * 20000 / sb.getConfigLong(SwitchboardConstants.WORDCACHE_MAX_COUNT, 100000)); // estimation of necessary pause time
}
* Complete number localization and provide a more reasonable interface to serverObjects: - put(key, value) methods are now used if a value added to the map should be kept as it is. Numbers are transformed (but not formatted) to an equivalent String representation. - putASIS(...) have been removed, now done with simple put(...) (see above). - puNum(...) can be used for number values which should be stored in a formatted way, either depending on the current locale setting for yacy (default) or in a "none" locale (see javadocs and setLocalize()). - putHTML(...) escapes special characters into corresponding HTML enities ('<' => '&lt;') which was done with put(...) before and so was called too often, becauses it is necessary only for very few cases. Additionally there is a "forXML" mode which only replaces < > & ". In short: Use put(...) for almost everything, use putXY(...) if you need some special transformation of the value. A few bugs have been fixed as well, and there should be a small performance improvement for complex pages with a lot of values. * added additional Sum/Avg rows to access tracker pages, see http://forum.yacy-websuche.de/viewtopic.php?f=5&t=456 * removed duplicate code (mostly related to the big changes above). TODO: - make sure, number formats work as expected _everywhere_, report overseen stuff http://forum.yacy-websuche.de/viewtopic.php?f=5&t=437 - probably a good idea to add special putDate() methods as they are used in many pages and create duplicated formatting code + maybe some centralized handling for memory value formatting. - further improve the speed of page creation for the WatchCrawler. git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@4178 6c8d7289-2bf4-0310-a012-ef5d649a1542
17 years ago
prop.put("unknownURL", unknownURLs.toString());
prop.put("result", result);
prop.put("pause", pause);
// return rewrite properties
return prop;
}
/**
* @param requestIdentifier
* @param msg
*/
private static void logWarning(final String requestIdentifier, final String msg) {
ConcurrentLog.warn("transferRWI", requestIdentifier +" "+ msg);
}
}