Add a cursor for tracking monthly uploads of ICANN report (#343)

* Add a cursor for tracking monthly uploads of the transaction report to ICANN

* Add cursors to track activity, transaction, and manifest report uploads.

* Address comments

* Add @Nullable annotation to manifestCursor

* Add lock and batch load cursors.

* Add string formatting, autovalue CursorInfo object, and handling for null cursors

* Add some helper functions for loadCursors and restructure to require less round trips to the database

* Switch new cursors to be created with cursorTime at first of next month
This commit is contained in:
sarahcaseybot
2019-11-22 17:40:31 -05:00
committed by GitHub
parent cc5f62587e
commit e318f47fc6
2 changed files with 441 additions and 99 deletions
@@ -15,34 +15,51 @@
package google.registry.reporting.icann;
import static com.google.common.base.Preconditions.checkArgument;
import static com.google.common.collect.ImmutableMap.toImmutableMap;
import static com.google.common.net.MediaType.PLAIN_TEXT_UTF_8;
import static google.registry.model.common.Cursor.getCursorTimeOrStartOfTime;
import static google.registry.model.ofy.ObjectifyService.ofy;
import static google.registry.model.transaction.TransactionManagerFactory.tm;
import static google.registry.reporting.icann.IcannReportingModule.MANIFEST_FILE_NAME;
import static google.registry.reporting.icann.IcannReportingModule.PARAM_SUBDIR;
import static google.registry.request.Action.Method.POST;
import static java.nio.charset.StandardCharsets.UTF_8;
import static javax.servlet.http.HttpServletResponse.SC_OK;
import com.google.appengine.tools.cloudstorage.GcsFilename;
import com.google.common.base.Splitter;
import com.google.common.collect.ImmutableList;
import com.google.auto.value.AutoValue;
import com.google.common.collect.ImmutableMap;
import com.google.common.collect.ImmutableSet;
import com.google.common.flogger.FluentLogger;
import com.google.common.io.ByteStreams;
import com.googlecode.objectify.Key;
import google.registry.config.RegistryConfig.Config;
import google.registry.gcs.GcsUtils;
import google.registry.model.common.Cursor;
import google.registry.model.common.Cursor.CursorType;
import google.registry.model.registry.Registries;
import google.registry.model.registry.Registry;
import google.registry.model.registry.Registry.TldType;
import google.registry.request.Action;
import google.registry.request.HttpException.ServiceUnavailableException;
import google.registry.request.Parameter;
import google.registry.request.Response;
import google.registry.request.auth.Auth;
import google.registry.request.lock.LockHandler;
import google.registry.util.Clock;
import google.registry.util.EmailMessage;
import google.registry.util.Retrier;
import google.registry.util.SendEmailService;
import java.io.IOException;
import java.io.InputStream;
import java.util.Map;
import java.util.concurrent.Callable;
import java.util.regex.Pattern;
import java.util.stream.Collectors;
import javax.annotation.Nullable;
import javax.inject.Inject;
import javax.mail.internet.InternetAddress;
import org.joda.time.DateTime;
import org.joda.time.Duration;
/**
* Action that uploads the monthly activity/transactions reports from GCS to ICANN via an HTTP PUT.
@@ -81,44 +98,175 @@ public final class IcannReportingUploadAction implements Runnable {
@Inject @Config("gSuiteOutgoingEmailAddress") InternetAddress sender;
@Inject @Config("alertRecipientEmailAddress") InternetAddress recipient;
@Inject SendEmailService emailService;
@Inject Clock clock;
@Inject LockHandler lockHandler;
@Inject
IcannReportingUploadAction() {}
@Override
public void run() {
String reportBucketname = String.format("%s/%s", reportingBucket, subdir);
ImmutableList<String> manifestedFiles = getManifestedFiles(reportBucketname);
ImmutableMap.Builder<String, Boolean> reportSummaryBuilder = new ImmutableMap.Builder<>();
// Report on all manifested files
for (String reportFilename : manifestedFiles) {
logger.atInfo().log(
"Reading ICANN report %s from bucket %s", reportFilename, reportBucketname);
final GcsFilename gcsFilename = new GcsFilename(reportBucketname, reportFilename);
verifyFileExists(gcsFilename);
boolean success = false;
try {
success =
retrier.callWithRetry(
() -> {
final byte[] payload = readBytesFromGcs(gcsFilename);
return icannReporter.send(payload, reportFilename);
},
IcannReportingUploadAction::isUploadFailureRetryable);
} catch (RuntimeException e) {
logger.atWarning().withCause(e).log("Upload to %s failed.", gcsFilename);
}
reportSummaryBuilder.put(reportFilename, success);
Callable<Void> lockRunner =
() -> {
ImmutableMap.Builder<String, Boolean> reportSummaryBuilder = new ImmutableMap.Builder<>();
ImmutableMap<Cursor, CursorInfo> cursors = loadCursors();
// If cursor time is before now, upload the corresponding report
cursors.entrySet().stream()
.filter(entry -> getCursorTimeOrStartOfTime(entry.getKey()).isBefore(clock.nowUtc()))
.forEach(
entry -> {
DateTime cursorTime = getCursorTimeOrStartOfTime(entry.getKey());
uploadReport(
cursorTime,
entry.getValue().getType(),
entry.getValue().getTld(),
reportSummaryBuilder);
});
// Send email of which reports were uploaded
emailUploadResults(reportSummaryBuilder.build());
response.setStatus(SC_OK);
response.setContentType(PLAIN_TEXT_UTF_8);
return null;
};
String lockname = "IcannReportingUploadAction";
if (!lockHandler.executeWithLocks(lockRunner, null, Duration.standardHours(2), lockname)) {
throw new ServiceUnavailableException("Lock for IcannReportingUploadAction already in use");
}
emailUploadResults(reportSummaryBuilder.build());
response.setStatus(SC_OK);
response.setContentType(PLAIN_TEXT_UTF_8);
response.setPayload(
String.format("OK, attempted uploading %d reports", manifestedFiles.size()));
}
/** Uploads the report and rolls forward the cursor for that report. */
private void uploadReport(
DateTime cursorTime,
CursorType cursorType,
String tldStr,
ImmutableMap.Builder<String, Boolean> reportSummaryBuilder) {
String reportBucketname = String.format("%s/%s", reportingBucket, subdir);
String filename = getFileName(cursorType, cursorTime, tldStr);
final GcsFilename gcsFilename = new GcsFilename(reportBucketname, filename);
logger.atInfo().log("Reading ICANN report %s from bucket %s", filename, reportBucketname);
// Check that the report exists
try {
verifyFileExists(gcsFilename);
} catch (IllegalArgumentException e) {
String logMessage =
String.format(
"Could not upload %s report for %s because file %s did not exist.",
cursorType, tldStr, filename);
if (clock.nowUtc().dayOfMonth().get() == 1) {
logger.atInfo().withCause(e).log(logMessage + " This report may not have been staged yet.");
} else {
logger.atSevere().withCause(e).log(logMessage);
}
reportSummaryBuilder.put(filename, false);
return;
}
// Upload the report
boolean success = false;
try {
success =
retrier.callWithRetry(
() -> {
final byte[] payload = readBytesFromGcs(gcsFilename);
return icannReporter.send(payload, filename);
},
IcannReportingUploadAction::isUploadFailureRetryable);
} catch (RuntimeException e) {
logger.atWarning().withCause(e).log("Upload to %s failed", gcsFilename);
}
reportSummaryBuilder.put(filename, success);
// Set cursor to first day of next month if the upload succeeded
if (success) {
Cursor newCursor;
if (cursorType.equals(CursorType.ICANN_UPLOAD_MANIFEST)) {
newCursor =
Cursor.createGlobal(
cursorType, cursorTime.withTimeAtStartOfDay().withDayOfMonth(1).plusMonths(1));
} else {
newCursor =
Cursor.create(
cursorType,
cursorTime.withTimeAtStartOfDay().withDayOfMonth(1).plusMonths(1),
Registry.get(tldStr));
}
tm().transact(() -> ofy().save().entity(newCursor));
}
}
private String getFileName(CursorType cursorType, DateTime cursorTime, String tld) {
if (cursorType.equals(CursorType.ICANN_UPLOAD_MANIFEST)) {
return MANIFEST_FILE_NAME;
}
return String.format(
"%s%s%d%02d.csv",
tld,
(cursorType.equals(CursorType.ICANN_UPLOAD_ACTIVITY) ? "-activity-" : "-transactions-"),
cursorTime.year().get(),
cursorTime.monthOfYear().get());
}
/** Returns a map of each cursor to the CursorType and tld. */
private ImmutableMap<Cursor, CursorInfo> loadCursors() {
ImmutableSet<Registry> registries = Registries.getTldEntitiesOfType(TldType.REAL);
Map<Key<Cursor>, Registry> activityKeyMap =
loadKeyMap(registries, CursorType.ICANN_UPLOAD_ACTIVITY);
Map<Key<Cursor>, Registry> transactionKeyMap =
loadKeyMap(registries, CursorType.ICANN_UPLOAD_TX);
ImmutableSet.Builder<Key<Cursor>> keys = new ImmutableSet.Builder<>();
keys.addAll(activityKeyMap.keySet());
keys.addAll(transactionKeyMap.keySet());
keys.add(Cursor.createGlobalKey(CursorType.ICANN_UPLOAD_MANIFEST));
Map<Key<Cursor>, Cursor> cursorMap = ofy().load().keys(keys.build());
ImmutableMap.Builder<Cursor, CursorInfo> cursors = new ImmutableMap.Builder<>();
defaultNullCursorsToNextMonthAndAddToMap(
activityKeyMap, CursorType.ICANN_UPLOAD_ACTIVITY, cursorMap, cursors);
defaultNullCursorsToNextMonthAndAddToMap(
transactionKeyMap, CursorType.ICANN_UPLOAD_TX, cursorMap, cursors);
Cursor manifestCursor =
cursorMap.getOrDefault(
Cursor.createGlobalKey(CursorType.ICANN_UPLOAD_MANIFEST),
Cursor.createGlobal(CursorType.ICANN_UPLOAD_MANIFEST, clock.nowUtc().minusDays(1)));
cursors.put(manifestCursor, CursorInfo.create(CursorType.ICANN_UPLOAD_MANIFEST, null));
return cursors.build();
}
private Map<Key<Cursor>, Registry> loadKeyMap(
ImmutableSet<Registry> registries, CursorType type) {
return registries.stream().collect(toImmutableMap(r -> Cursor.createKey(type, r), r -> r));
}
/**
* Populate the cursors map with the Cursor and CursorInfo for each key in the keyMap. If the key
* from the keyMap does not have an existing cursor, create a new cursor with a default cursorTime
* of the first of next month.
*/
private void defaultNullCursorsToNextMonthAndAddToMap(
Map<Key<Cursor>, Registry> keyMap,
CursorType type,
Map<Key<Cursor>, Cursor> cursorMap,
ImmutableMap.Builder<Cursor, CursorInfo> cursors) {
keyMap.forEach(
(key, registry) -> {
// Cursor time is defaulted to the first of next month since a new tld will not yet have a
// report staged for upload.
Cursor cursor =
cursorMap.getOrDefault(
key, Cursor.create(type, clock.nowUtc().minusDays(1), registry));
cursors.put(cursor, CursorInfo.create(type, registry.getTldStr()));
});
}
/** Don't retry when reports are already uploaded or can't be uploaded. */
private static final String ICANN_UPLOAD_PERMANENT_ERROR_MESSAGE =
"A report for that month already exists, the cut-off date already passed.";
"A report for that month already exists, the cut-off date already passed";
/** Don't retry when the IP address isn't whitelisted, as retries go through the same IP. */
private static final Pattern ICANN_UPLOAD_WHITELIST_ERROR =
@@ -146,18 +294,6 @@ public final class IcannReportingUploadAction implements Runnable {
emailService.sendEmail(EmailMessage.create(subject, body, recipient, sender));
}
private ImmutableList<String> getManifestedFiles(String reportBucketname) {
GcsFilename manifestFilename = new GcsFilename(reportBucketname, MANIFEST_FILE_NAME);
verifyFileExists(manifestFilename);
return retrier.callWithRetry(
() ->
ImmutableList.copyOf(
Splitter.on('\n')
.omitEmptyStrings()
.split(new String(readBytesFromGcs(manifestFilename), UTF_8))),
IOException.class);
}
private byte[] readBytesFromGcs(GcsFilename reportFilename) throws IOException {
try (InputStream gcsInput = gcsUtils.openInputStream(reportFilename)) {
return ByteStreams.toByteArray(gcsInput);
@@ -171,4 +307,16 @@ public final class IcannReportingUploadAction implements Runnable {
gcsFilename.getObjectName(),
gcsFilename.getBucketName());
}
@AutoValue
abstract static class CursorInfo {
static CursorInfo create(CursorType type, @Nullable String tld) {
return new AutoValue_IcannReportingUploadAction_CursorInfo(type, tld);
}
public abstract CursorType getType();
@Nullable
abstract String getTld();
}
}