use of com.hubspot.singularity.api.SingularityS3SearchRequest in project Singularity by HubSpot.
the class S3LogResource method getS3LogsWithExecutorService.
// Fetching logs
private List<SingularityS3LogMetadata> getS3LogsWithExecutorService(S3Configuration s3Configuration, ListeningExecutorService executorService, Map<SingularityS3Service, Set<String>> servicesToPrefixes, int totalPrefixCount, final SingularityS3SearchRequest search, final ConcurrentHashMap<String, ContinuationToken> continuationTokens, final boolean paginated) throws InterruptedException, ExecutionException, TimeoutException {
List<ListenableFuture<List<S3ObjectSummaryHolder>>> futures = Lists.newArrayListWithCapacity(totalPrefixCount);
final AtomicInteger resultCount = new AtomicInteger();
for (final Map.Entry<SingularityS3Service, Set<String>> entry : servicesToPrefixes.entrySet()) {
final String s3Bucket = entry.getKey().getBucket();
final String group = entry.getKey().getGroup();
final AmazonS3 s3Client = entry.getKey().getS3Client();
for (final String s3Prefix : entry.getValue()) {
final String key = String.format(CONTINUATION_TOKEN_KEY_FORMAT, group, s3Bucket, s3Prefix);
if (search.getContinuationTokens().containsKey(key) && search.getContinuationTokens().get(key).isLastPage()) {
LOG.trace("No further content for prefix {} in bucket {}, skipping", s3Prefix, s3Bucket);
continuationTokens.putIfAbsent(key, search.getContinuationTokens().get(key));
continue;
}
futures.add(executorService.submit(new Callable<List<S3ObjectSummaryHolder>>() {
@Override
public List<S3ObjectSummaryHolder> call() throws Exception {
ListObjectsV2Request request = new ListObjectsV2Request().withBucketName(s3Bucket).withPrefix(s3Prefix);
if (paginated) {
Optional<ContinuationToken> token = Optional.empty();
if (search.getContinuationTokens().containsKey(key) && !Strings.isNullOrEmpty(search.getContinuationTokens().get(key).getValue())) {
request.setContinuationToken(search.getContinuationTokens().get(key).getValue());
token = Optional.of(search.getContinuationTokens().get(key));
}
int targetResultCount = search.getMaxPerPage().orElse(DEFAULT_TARGET_MAX_RESULTS);
request.setMaxKeys(targetResultCount);
if (resultCount.get() < targetResultCount) {
ListObjectsV2Result result = s3Client.listObjectsV2(request);
if (result.getObjectSummaries().isEmpty()) {
continuationTokens.putIfAbsent(key, new ContinuationToken(result.getNextContinuationToken(), true));
return Collections.emptyList();
} else {
boolean addToList = incrementIfLessThan(resultCount, result.getObjectSummaries().size(), targetResultCount);
if (addToList) {
continuationTokens.putIfAbsent(key, new ContinuationToken(result.getNextContinuationToken(), !result.isTruncated()));
List<S3ObjectSummaryHolder> objectSummaryHolders = new ArrayList<>();
for (S3ObjectSummary objectSummary : result.getObjectSummaries()) {
if (search.getFileNamePrefixWhitelist().isEmpty() || search.getFileNamePrefixWhitelist().stream().anyMatch(whitelistedPrefix -> objectSummary.getKey().startsWith(whitelistedPrefix))) {
objectSummaryHolders.add(new S3ObjectSummaryHolder(group, objectSummary));
}
}
return objectSummaryHolders;
} else {
continuationTokens.putIfAbsent(key, token.orElse(new ContinuationToken(null, false)));
return Collections.emptyList();
}
}
} else {
continuationTokens.putIfAbsent(key, token.orElse(new ContinuationToken(null, false)));
return Collections.emptyList();
}
} else {
ListObjectsV2Result result = s3Client.listObjectsV2(request);
List<S3ObjectSummaryHolder> objectSummaryHolders = new ArrayList<>();
for (S3ObjectSummary objectSummary : result.getObjectSummaries()) {
if (search.getFileNamePrefixWhitelist().isEmpty() || search.getFileNamePrefixWhitelist().stream().anyMatch(whitelistedPrefix -> objectSummary.getKey().startsWith(whitelistedPrefix))) {
objectSummaryHolders.add(new S3ObjectSummaryHolder(group, objectSummary));
}
}
while (result.isTruncated() && result.getContinuationToken() != null) {
result = s3Client.listObjectsV2(new ListObjectsV2Request().withBucketName(s3Bucket).withPrefix(s3Prefix).withContinuationToken(result.getContinuationToken()));
for (S3ObjectSummary objectSummary : result.getObjectSummaries()) {
objectSummaryHolders.add(new S3ObjectSummaryHolder(group, objectSummary));
}
}
return objectSummaryHolders;
}
}
}));
}
}
final long start = System.currentTimeMillis();
List<List<S3ObjectSummaryHolder>> results = Futures.allAsList(futures).get(s3Configuration.getWaitForS3ListSeconds(), TimeUnit.SECONDS);
List<S3ObjectSummaryHolder> objects = Lists.newArrayListWithExpectedSize(results.size() * 2);
for (List<S3ObjectSummaryHolder> s3ObjectSummaryHolders : results) {
for (final S3ObjectSummaryHolder s3ObjectHolder : s3ObjectSummaryHolders) {
objects.add(s3ObjectHolder);
}
}
LOG.trace("Got {} objects from S3 after {}", objects.size(), JavaUtils.duration(start));
List<ListenableFuture<SingularityS3LogMetadata>> logFutures = Lists.newArrayListWithCapacity(objects.size());
final Date expireAt = new Date(System.currentTimeMillis() + s3Configuration.getExpireS3LinksAfterMillis());
for (final S3ObjectSummaryHolder s3ObjectHolder : objects) {
final S3ObjectSummary s3Object = s3ObjectHolder.getObjectSummary();
final AmazonS3 s3Client = s3Services.getServiceByGroupAndBucketOrDefault(s3ObjectHolder.getGroup(), s3Object.getBucketName()).getS3Client();
logFutures.add(executorService.submit(new Callable<SingularityS3LogMetadata>() {
@Override
public SingularityS3LogMetadata call() throws Exception {
Optional<Long> maybeStartTime = Optional.empty();
Optional<Long> maybeEndTime = Optional.empty();
if (!search.isExcludeMetadata()) {
GetObjectMetadataRequest metadataRequest = new GetObjectMetadataRequest(s3Object.getBucketName(), s3Object.getKey());
Map<String, String> objectMetadata = s3Client.getObjectMetadata(metadataRequest).getUserMetadata();
maybeStartTime = getMetadataAsLong(objectMetadata, SingularityS3Log.LOG_START_S3_ATTR);
maybeEndTime = getMetadataAsLong(objectMetadata, SingularityS3Log.LOG_END_S3_ATTR);
}
if (search.isListOnly()) {
return new SingularityS3LogMetadata(s3Object.getKey(), s3Object.getLastModified().getTime(), s3Object.getSize(), maybeStartTime, maybeEndTime);
} else {
GeneratePresignedUrlRequest getUrlRequest = new GeneratePresignedUrlRequest(s3Object.getBucketName(), s3Object.getKey()).withMethod(HttpMethod.GET).withExpiration(expireAt);
String getUrl = s3Client.generatePresignedUrl(getUrlRequest).toString();
ResponseHeaderOverrides downloadHeaders = new ResponseHeaderOverrides();
downloadHeaders.setContentDisposition(CONTENT_DISPOSITION_DOWNLOAD_HEADER);
downloadHeaders.setContentEncoding(CONTENT_ENCODING_DOWNLOAD_HEADER);
GeneratePresignedUrlRequest downloadUrlRequest = new GeneratePresignedUrlRequest(s3Object.getBucketName(), s3Object.getKey()).withMethod(HttpMethod.GET).withExpiration(expireAt).withResponseHeaders(downloadHeaders);
String downloadUrl = s3Client.generatePresignedUrl(downloadUrlRequest).toString();
return new SingularityS3Log(getUrl, s3Object.getKey(), s3Object.getLastModified().getTime(), s3Object.getSize(), downloadUrl, maybeStartTime, maybeEndTime);
}
}
}));
}
return Futures.allAsList(logFutures).get(s3Configuration.getWaitForS3LinksSeconds(), TimeUnit.SECONDS);
}
use of com.hubspot.singularity.api.SingularityS3SearchRequest in project Singularity by HubSpot.
the class S3LogResource method getS3LogsForTask.
@GET
@Path("/task/{taskId}")
@Operation(summary = "Retrieve the list of logs stored in S3 for a specific task", responses = { @ApiResponse(responseCode = "200", description = "Returns a list of metadata about log files for the specified task", content = { @Content(array = @ArraySchema(schema = @Schema(implementation = SingularityS3LogMetadata.class))), @Content(array = @ArraySchema(schema = @Schema(implementation = SingularityS3Log.class))) }), @ApiResponse(responseCode = "404", description = "S3 configuration is not present") })
public List<SingularityS3LogMetadata> getS3LogsForTask(@Parameter(hidden = true) @Auth SingularityUser user, @Parameter(required = true, description = "The task ID to search for") @PathParam("taskId") String taskId, @Parameter(description = "Start timestamp (millis, 13 digit)") @QueryParam("start") Optional<Long> start, @Parameter(description = "End timestamp (mills, 13 digit)") @QueryParam("end") Optional<Long> end, @Parameter(description = "Exclude custom object metadata") @QueryParam("excludeMetadata") @DefaultValue("false") boolean excludeMetadata, @Parameter(description = "Do not generate download/get urls, only list the files and metadata") @QueryParam("list") @DefaultValue("false") boolean listOnly) throws Exception {
checkS3();
final SingularityS3SearchRequest search = new SingularityS3SearchRequest(Collections.emptyMap(), Collections.singletonList(taskId), start, end, excludeMetadata, listOnly, Optional.empty(), Collections.emptyMap());
try {
return getS3Logs(configuration.get(), getServiceToPrefixes(search, user), search, false).getResults();
} catch (TimeoutException te) {
throw timeout("Timed out waiting for response from S3 for %s", taskId);
} catch (Throwable t) {
throw new RuntimeException(t);
}
}
Aggregations