Error : Output file full, and no more output files

[bskumar4u]

I have a job which just funnels about 1billion records from a .ds and around 202million records from another .ds
When the job is running for one and half hour(around900M rec), its getting aborted with the following error.
Note: Funnel is followed by a sort.
As per my knowledge this type of error usually occurs when system runs out of virtual(buffering) memory that buffers the data.
i mean in case of join until the match is found
incase of sort until it receives all the records before peroming sort.
please correct my understanding and please do let me know how can we overcome this issue.
Note: the amount of data may increase further upto 2billion.

APT_CombinedOperatorController(1),1: write failed: Output file full, and no more output files
APT_CombinedOperatorController(1),1: Failure during execution of operator logic.
APT_CombinedOperatorController(1),1: Fatal Error: Tsort merger aborting: mergeOneRecord() punted
subArgs={asc}}}(0),1: Failure during execution of operator logic.
subArgs={asc}}}(0),1: Input 0 consumed 325493715 records.
subArgs={asc}}}(0),1: Output 0 produced 0 records.
APT_CombinedOperatorController(1),1: Fatal Error: Pipe read failed: short read
buffer(3),1: Failure during execution of operator logic.
.....
.
.
.
.
.

[qt_ky]

It sounds like you need to well more-than-double your disk space to allow for large sorts. Open the config file used by the job, or just look for it in the job log details. Find the resource scratchdisk entries per node. Then go out to your operating system and check the disk space for those locations. Clean up any old files that may be left around from past aborts. And add more disk space.