Jobs aborting and universe.log displays error “Unable to write local file”

Document ID : KB000087103
Last Modified Date : 14/04/2018
Show Technical Document Details
Issue:
Error Message :
In universe.log:
|ERROR|X|END|pid=16921.4151483184| u_write_file | Error fopen for writing
|ERROR|X|END|pid=16921.4151483184| decodeFileResponse | Unable to write local file [/orsyp/DUAS/COMPANY/temp/U_TMP_0003] (u_write_file returns 1807)

Patch level detected:Dollar Universe 6.2.01
The node is not able to submit jobs.

When you check the universe.log errors similar to these are present:

 
| 2016-06-02 17:51:06 |ERROR|X|IO |pid=5385.117283696| u_write_file | Error fopen for writing
| 2016-06-02 17:51:06 |ERROR|X|IO |pid=5385.117283696| decodeFileResponse | Unable to write local file [/orsyp/DUAS/COMPANY/temp/U_TMP_0003] (u_write_file returns 1807)
| 2016-06-02 17:51:06 |ERROR|X|IO |pid=5385.117283696| u_ft_callsrv_r | u_ntoh_reponse_file returns error

If you try to do an UxTrace or launch an offline reorganization, both procedures are aborted.
 
Environment:
OS: All
OS Version: N/A
Cause:
Cause type:
Configuration
Root Cause: The error “Unable to write local file [/orsyp/DUAS/COMPANY/temp/U_TMP_0003]” means that Dollar Universe cannot write in temp directory because is full.
The temp directory is used during the following actions:
- Normally during the execution of a job, a copy of the job is created here, when the job finish successfully the temporary file is deleted.
- During the offline reorganization a copy of the data file is created here, when the reorganization finish successfully the temporary file is deleted.
- During the UxTrace procedure

Dollar Universe temp directory can be full for some reasons:
-When one job is aborted the temp directory will keep the copy created, this means that after a crash of Dollar Universe so much temporary files will be created provoking a file system full
- If one reorganization is stopped, after the relaunch of Dollar Universe the temporary file created will remain in temp directory and provokes a file system full
-If a reorganization offline is launched but there is not enough space to copy .dta file
- If an UxTrace procedure is launched but there is not enough space to copy the files
Resolution:
Increase the filesystem hosting the temp folder and/or delete unnecessary files from the filesystem.

-In order to keep the temp directory from filling up, the copy of jobs in aborted status should be removed manually.  Add this step
to your maintenance procedure.

- After a crash, check to ensure that the temp directory 
has enough space free.  if it doesn't then purge the directory manually.

-If an offline reorganization is interrupted, check that there are no temporary file remaining.

-Before launching the
UxTrace or an offline reorganization, ensure that enough space is free to copy your data files.
 

Fix Status: No Fix

Fix Version(s):
N/A
Additional Information:
Workaround :
N/A