s3 storage: NIO S3 library does not support glob

[buchanae]

In short, while using the S3 backend with TES, I realized the cwl.output.json was not being created. An uncaught exception was silently killing the server. The exception occurs because the S3 library doesn't support glob.

This line:
https://github.com/rabix/bunny/blob/develop/rabix-bindings-cwl/src/main/java/org/rabix/bindings/cwl/service/impl/CWLGlobServiceImpl.java#L57

Results in an uncaught exception, which maybe should be caught here:
https://github.com/rabix/bunny/blob/develop/rabix-bindings-cwl/src/main/java/org/rabix/bindings/cwl/CWLProcessor.java#L150

The stacktrace:

15:08:38.923 [pool-2-thread-1] ERROR org.rabix.bindings.cwl.CWLProcessor - Uncaught exception 
java.lang.UnsupportedOperationException: null
	at com.upplication.s3fs.S3FileSystem.getPathMatcher(S3FileSystem.java:105) ~[Amazon-S3-FileSystem-NIO2-b3456a1692f82d500c302f9e33ab803d9348bbc6.jar:na]
	at org.rabix.bindings.cwl.service.impl.CWLGlobServiceImpl.glob(CWLGlobServiceImpl.java:57) ~[classes/:na]
	at org.rabix.bindings.cwl.CWLProcessor.globFiles(CWLProcessor.java:380) ~[classes/:na]
	at org.rabix.bindings.cwl.CWLProcessor.collectOutput(CWLProcessor.java:316) ~[classes/:na]
	at org.rabix.bindings.cwl.CWLProcessor.collectOutputs(CWLProcessor.java:178) ~[classes/:na]
	at org.rabix.bindings.cwl.CWLProcessor.postprocess(CWLProcessor.java:147) ~[classes/:na]
	at org.rabix.bindings.cwl.CWLBindings.postprocess(CWLBindings.java:198) [classes/:na]
	at org.rabix.backend.tes.service.impl.LocalTESWorkerServiceImpl.success(LocalTESWorkerServiceImpl.java:119) [classes/:na]
	at org.rabix.backend.tes.service.impl.LocalTESWorkerServiceImpl.access$100(LocalTESWorkerServiceImpl.java:78) [classes/:na]
	at org.rabix.backend.tes.service.impl.LocalTESWorkerServiceImpl$1.run(LocalTESWorkerServiceImpl.java:168) [classes/:na]
	at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) [na:1.8.0_102]
	at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:308) [na:1.8.0_102]
	at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:180) [na:1.8.0_102]
	at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:294) [na:1.8.0_102]
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) [na:1.8.0_102]
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) [na:1.8.0_102]
	at java.lang.Thread.run(Thread.java:745) [na:1.8.0_102]
15:08:38.924 [pool-2-thread-1] ERROR o.r.b.t.s.i.LocalTESWorkerServiceImpl - Couldn't process job
org.rabix.bindings.BindingException: java.lang.UnsupportedOperationException
	at org.rabix.bindings.cwl.CWLProcessor.postprocess(CWLProcessor.java:154) ~[classes/:na]
	at org.rabix.bindings.cwl.CWLBindings.postprocess(CWLBindings.java:198) ~[classes/:na]
	at org.rabix.backend.tes.service.impl.LocalTESWorkerServiceImpl.success(LocalTESWorkerServiceImpl.java:119) [classes/:na]
	at org.rabix.backend.tes.service.impl.LocalTESWorkerServiceImpl.access$100(LocalTESWorkerServiceImpl.java:78) [classes/:na]
	at org.rabix.backend.tes.service.impl.LocalTESWorkerServiceImpl$1.run(LocalTESWorkerServiceImpl.java:168) [classes/:na]
	at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) [na:1.8.0_102]
	at java.util.concurrent.FutureTask.runAndReset(FutureTask.java:308) [na:1.8.0_102]
	at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:180) [na:1.8.0_102]
	at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:294) [na:1.8.0_102]
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) [na:1.8.0_102]
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) [na:1.8.0_102]
	at java.lang.Thread.run(Thread.java:745) [na:1.8.0_102]
Caused by: java.lang.UnsupportedOperationException: null
	at com.upplication.s3fs.S3FileSystem.getPathMatcher(S3FileSystem.java:105) ~[Amazon-S3-FileSystem-NIO2-b3456a1692f82d500c302f9e33ab803d9348bbc6.jar:na]
	at org.rabix.bindings.cwl.service.impl.CWLGlobServiceImpl.glob(CWLGlobServiceImpl.java:57) ~[classes/:na]
	at org.rabix.bindings.cwl.CWLProcessor.globFiles(CWLProcessor.java:380) ~[classes/:na]
	at org.rabix.bindings.cwl.CWLProcessor.collectOutput(CWLProcessor.java:316) ~[classes/:na]
	at org.rabix.bindings.cwl.CWLProcessor.collectOutputs(CWLProcessor.java:178) ~[classes/:na]
	at org.rabix.bindings.cwl.CWLProcessor.postprocess(CWLProcessor.java:147) ~[classes/:na]
	... 11 common frames omitted

[buchanae]

The method that needs to be implemented is here:
https://github.com/Upplication/Amazon-S3-FileSystem-NIO2/blob/master/src/main/java/com/upplication/s3fs/S3FileSystem.java#L102

I'm thinking of implementing a wrapper around that library that implements that method for globs on s3. I can't find any other libraries for NIO + S3. Thoughts?

[buchanae]

I tried to implement a wrapper for this today, but got stuck at the NIO file system provider magic. Apparently it loads filesystem implementations from the classpath + JAR files, meaning it's not straightforward to implement a wrapper. At least, it's beyond my java skills for now.

With a little advice on how to deal with that issue, I can help implement glob.

[milos-ljubinkovic]

The library supports file walking, it's just that the path matcher isn't implemented, but using a regular UnixPath matcher works just as well. I've made this change in the latest release.