Cryptic R errors need better reporting

[lhsego]

This error, which I produced using datadr::makeDisplay() on an hdfsConn ddo,

---------------------------------
There were R errors, showing 30:

Warning message:
Autokill is true and terminating job_1441994449703_0070
Error in .jcall("RJavaTools", "Ljava/lang/Object;", "invokeMethod", cl,  : 
  java.io.FileNotFoundException: Cannot access /user/d3p423/tmp/tmp_output-6ae2a6d4f0966b3244a47a8f588f6c36: No such file or directory.
Calls: makeDisplay ... <Anonymous> -> .jrcall -> .jcall -> .jcheck -> .Call
In addition: Warning message:
In Rhipe:::rhwatch.runner(job = job, mon.sec = mon.sec, readback = readback,  :
 Job failure, deleting output: /user/d3p423/tmp/tmp_output-6ae2a6d4f0966b3244a47a8f588f6c36:   
Execution halted 
Warning message:
system call failed: Cannot allocate memory 

is apparently a red-herring when there are R errors on the hadoop job--but the only error that is well described is the Java error not being able to access a file. Would sure be nice to have a sense of what the R errors are.

[saptarshiguha]

Hmm, true. I have a feeling that you'll find the R error in the job log and
for some reason it didn't make it's way back to the R console.

On Mon, Oct 19, 2015 at 7:50 PM, Landon Sego [email protected]
wrote:

This error, which I produced using datadr::makeDisplay() on an hdfsConn
ddo,

---

There were R errors, showing 30:

Warning message:
Autokill is true and terminating job_1441994449703_0070
Error in .jcall("RJavaTools", "Ljava/lang/Object;", "invokeMethod", cl, :
java.io.FileNotFoundException: Cannot access /user/d3p423/tmp/tmp_output-6ae2a6d4f0966b3244a47a8f588f6c36: No such file or directory.
Calls: makeDisplay ... -> .jrcall -> .jcall -> .jcheck -> .Call
In addition: Warning message:
In Rhipe:::rhwatch.runner(job = job, mon.sec = mon.sec, readback = readback, :
Job failure, deleting output: /user/d3p423/tmp/tmp_output-6ae2a6d4f0966b3244a47a8f588f6c36:
Execution halted
Warning message:
system call failed: Cannot allocate memory

is apparently a red-herring when there are R errors on the hadoop job--but
the only error that is well described is the Java error not being able to
access a file. Would sure be nice to have a sense of what the R errors are.

—
Reply to this email directly or view it on GitHub
#32.