IBM Mainframe Forum

Hi All,

Some jobs in one of our application fails every now and then due to lack of space in few datasets.

Is there a way I can get a notification in case the space in these datasets goes beyond a particular threshold value(say 80%).
I have been told that the mainframe storage team can get such notification. I am looking at something that I from the application support team can develop and use to alert me in case the space in a dataset is almost full without having to depend on the mainframe storage team.

Thanks in advance.

Regards,
Ashok Kumar

Talk to your site storage group team about how to get the alerts you need. Developing your own solution would be wasteful of resources, and there is no reason to duplicate the work your storage group has already done.

Hi Robert,

Thanks for the response. I checked with them and they informed me that they are using a tool provided by a 3rd party vendor for such notification. They also have a limitation on the number of dataset that we can add to such alert from each application.

Is there anything you are aware of that can help me develop such utility?

if the failure happens in mission critical application
there is no reason for Your organization to use for Your application
the facilities provided by the third party tool.

do not expect too much help when You are facing issues due to Your organization <political> choices

It is not a good idea to develop your own system tools. If you still think you need to do this, you might be able to use DCOLLECT (if you have authority at your site to run DCOLLECT) to pull out allocation and used space for each data set on disk.

These users thanked the author Robert Sample for the post:

ashokkumar_rn (Tue Jun 25, 2013 3:10 pm)

Thanks Robert.

I tried using DCOLLECT but as suspected I do not have the authority to use it.

I tried using another utility IEHLIST which I found in some forums. The JCL I used for this is as shown below -

//LISTVTOC EXEC PGM=IEHLIST
//SYSPRINT DD SYSOUT=*
//DISK DD UNIT=3390,VOL=SER=KTSO12,DISP=SHR
//SYSIN DD *
LISTVTOC FORMAT,VOL=3390=KTSO12
/*

For each dataset in the volume KTS012 this JCL did display some information and one of the lines for this was as shown below -

"ON THE ABOVE DATA SET,THERE ARE 0 EMPTY CYLINDER(S) PLUS 11 EMPTY TRACK(S)."

Can the above statement help us in determining if the dataset is going out of space?

For eg: if the above line said 0 empty cylinders and 0 empty tracks then does this indicate this dataset may go out of space in near future?

That all depends on the size of the dataset and the amount of change it has. A 1 track dataset which has 11 tracks spare is not going to have any trouble unless you are expecting a gazillion records for it next time. Likewise a 1000 cyl dataset will be quite happy with 11 spare tracks if it never changes.

These users thanked the author NicC for the post:

ashokkumar_rn (Mon Jun 24, 2013 6:04 pm)

Hello,

Also keep in mind that VSAM files and physical database datasets will show "full" but mostly are not. The entire space is allocated and initialized, so it all appears "used".

Several of my clients have programmer-friendly datasets that contains dcollect data and smf data. Ask your storage management people if such is available on your system.

These users thanked the author dick scherrer for the post:

ashokkumar_rn (Tue Jun 25, 2013 3:10 pm)

Collecting allocated space and used space for sequential and "regular" partitioned data sets from the VTOC is not difficult. You are most of the way there using IEHLIST LISTVTOC FORMAT, though, as you have already determined, interpreting the output is not so easy.

As both NicC and Mr. Scherrer have mentioned, this analysis leaves many missing elements.

• VSAM
• Data bases
• Data sets migrated by HSM or ISV provided utilitiies (CA-Disk, for example) that perform similar functions.
• PDS-E data sets
• Data sets created during job execution.
• Tape data sets

Even when allocated and used space is available, determining if there is a potential problem can be difficult.

These users thanked the author steve-myers for the post:

ashokkumar_rn (Tue Jun 25, 2013 3:10 pm)

Thanks Steve,Dick,NicC,Robert and Enrico for your responses. The information you all have provided is very useful. I will try to develop something out of this and if I am able to get it right I will share it with this forum.

Thanks again.