- assigned_to: imi
Hi,
We notice this issue that occurs occasionally that cause the workflow to fail:
Gram Error Code: 17Error:
Server refused performing the request. Custom message: Server refused
creating directory (error code 1) [Nested exception message: Custom
message: Unexpected reply: 500-Command failed : System error in mkdir:
File exists
500-A system call failed: File exists
500 End.]
The error seems to be a gUSE issue. What gUSE normally does is
1) mkdir {a random folder - e.g. 0fbce168-011c-4e83-977c-128a7017ed90} via gridftp lib
2) Using gridftp lib, stage in wrapper.sh and localinputs.tgz
3) Submit to gram a job with wrapper.sh as an executable.
Unfortunately, it seemed that gUSE thought that step 1 had failed (maybe a timeout) and thus attempted to perform it again. Since step 1 actually worked (folder was successfully created), running the mkdir command the second time caused an error. The job then failed.
Suggestion:
Can you perhaps perform an additional check if the folder already exists before deciding to rerun step 1?
Best rgds,
Cerlane