search cancel

DCI Controller region gets a IOEZ00078E because zFS is getting full


Article ID: 134184


Updated On:


Dynamic Capacity Intelligence


The DCI Controller region gets an error message like the following one:

IOEZ00078E zFS aggregate dci.hql.zfs exceeds nn% full (ccccc/ttttt)

What is the best solution for this problem? 


Release : 1.5 or 2.0



The zFS cluster used for the Dynamic Capacity Intelligence (DCI) USS filesystem is getting full.


Two possible solutions: 

The quickest one, which doesn't require to shut down the Controller region, is to manually delete the old CSV files which are no longer used. These files can be found on the csv directory of the DCI USS filesystem and their names start with:


where yy is the last two digits of the year and mm is the month.

This will free some space on the zFS cluster and the files that were deleted can be recreated at a later time.
In addition, two other files need to be edited to eliminate the not working drop-down option in UI. It's WXR_Index.csv and Key_Info_300/900/3600.csv's. 
IN Key_info_300/900/3600.csv, you will look to the entries that starts with 'FIL', and based on what month and year CSV was deleted you should see the corresponding entry that should be deleted. Each LAPR has its own entry.  See screenhost below


In WXR_Index.csv. The highlighted part should helps you to decide which entries might be deleted, it's date, LPAR and Key_info file.  


The second solution requires that the Controller region could be stopped for a while, so the current zFS cluster can be copied on a larger one.

The steps required are the following ones:

  1. shut down the DCI Controller region
  2. unmount the zFS cluster
  3. increase the size of the cluster using the JCL contained on DCI 2.0 samplib member #DCIZFSC (attached to this article too)
  4. mount the new zFS cluster
  5. recreate the CSV files that were previously deleted (optional)
  6. restart the DCI Controller region


1562343699918__extend ZFS.txt get_app