Major Description
It is possible to configure S3 consumer to process several s3 objects in a single poll using the maxMessagesPerPoll property.
If this property contains a small number, less than 50, everything works fine but if user tries to consume more files then s3 consumer simply fails all the time. It cannot poll files because there are not enough HTTP connections to open streams for all the requested files at once. The exception looks like this:
com.amazonaws.AmazonClientException: Unable to execute HTTP request: Timeout waiting for connection from pool at com.amazonaws.http.AmazonHttpClient.executeHelper(AmazonHttpClient.java:544) at com.amazonaws.http.AmazonHttpClient.execute(AmazonHttpClient.java:273) at com.amazonaws.services.s3.AmazonS3Client.invoke(AmazonS3Client.java:3660) at com.amazonaws.services.s3.AmazonS3Client.getObject(AmazonS3Client.java:1133) at com.amazonaws.services.s3.AmazonS3EncryptionClient.access$201(AmazonS3EncryptionClient.java:65) at com.amazonaws.services.s3.AmazonS3EncryptionClient$S3DirectImpl.getObject(AmazonS3EncryptionClient.java:524) at com.amazonaws.services.s3.internal.crypto.S3CryptoModuleAE.getObjectSecurely(S3CryptoModuleAE.java:106) at com.amazonaws.services.s3.internal.crypto.CryptoModuleDispatcher.getObjectSecurely(CryptoModuleDispatcher.java:114) at com.amazonaws.services.s3.AmazonS3EncryptionClient.getObject(AmazonS3EncryptionClient.java:427) at com.amazonaws.services.s3.AmazonS3Client.getObject(AmazonS3Client.java:1005) at org.apache.camel.component.aws.s3.S3Consumer.createExchanges(S3Consumer.java:112) at org.apache.camel.component.aws.s3.S3Consumer.poll(S3Consumer.java:93) at org.apache.camel.impl.ScheduledPollConsumer.doRun(ScheduledPollConsumer.java:187) at org.apache.camel.impl.ScheduledPollConsumer.run(ScheduledPollConsumer.java:114) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:308) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:180) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:294) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) at java.lang.Thread.run(Thread.java:745)
The issue happens because by default AmazonS3Client uses HTTP client with limited number of connections in the pool - 50.
Since S3 consumer provides a possibility to consume any number of s3 objects in a single pool and because it is quite common case that someone needs to process 50 or more files in a single pool I think s3 consumer should handle this case properly. It should automatically change HTTP connections pool size to be able to handle requested number of objects. This can be done like this:
ClientConfiguration s3Config = new ClientConfiguration(); /* +20 we need to allocate a bit more to be sure that we always can do additional API calls when we already hold maxMessagesPerPoll s3 object streams opened */ s3Config.setMaxConnections(maxMessagesPerPoll + 20); AmazonS3Client client = new AeAmazonS3Client(awsCreds, s3Config );