View previous topic :: View next topic
|
Author |
Message |
ashek15
New User
Joined: 20 Oct 2015 Posts: 48 Location: USA
|
|
|
|
Dear All
Is there a way to check how much actual size was allocated to a sequential dataset which is used in a jcl which ran. Lets say for example allocation was SPACE=(CYL,(200,300),RLSE). Then Is there a way to know that out of total allocated size how much was actually allocated/used during jcl run. Just to mention the dataset is deleted after every run so we can't check actual allocated size in ispf 3.4 option.
Thanks |
|
Back to top |
|
|
Willy Jensen
Active Member
Joined: 01 Sep 2015 Posts: 712 Location: Denmark
|
|
|
|
Not directly as far as I am aware. DCOLLECT dataset data have some information. The SMFDCBBL of SMF type 15 is the count of blocks written, as I read it, and with a known blocksize you can calculate the number of tracks. |
|
Back to top |
|
|
vasanthz
Global Moderator
Joined: 28 Aug 2007 Posts: 1742 Location: Tirupur, India
|
|
|
|
I believe DCOLLECT cannot get that information as the dataset is deleted at the end of the JCL.
If DCOLLECT runs at the exact moment the dataset was allocated to its fullest size, then it would be able to get the information. Which is highly unlikely.
Also SMF 15 does not have the allocated space of a dataset, it does have primary and secondary space, but not the actual allocated space.
As for the answer to ashek15, I don't know :-( |
|
Back to top |
|
|
Rohit Umarjikar
Global Moderator
Joined: 21 Sep 2010 Posts: 3049 Location: NYC,USA
|
|
|
|
May be someone could help your better, if you tell us why you need that anyways even after DS is deleted? |
|
Back to top |
|
|
Robert Sample
Global Moderator
Joined: 06 Jun 2008 Posts: 8696 Location: Dubuque, Iowa, USA
|
|
|
|
The easiest way to collect this information is to have the application program write out the record count as part of its final tasks. |
|
Back to top |
|
|
ashek15
New User
Joined: 20 Oct 2015 Posts: 48 Location: USA
|
|
|
|
Thanks for your replies. To answer Rohit's query, we need this info to ensure we can track/monitor the sequential dataset usage and proactively allocate the appropriate space so that any future job failures can be avoided due to B37, D37 etc. |
|
Back to top |
|
|
Willy Jensen
Active Member
Joined: 01 Sep 2015 Posts: 712 Location: Denmark
|
|
|
|
@vasanthz - re 'I believe DCOLLECT cannot get that information as the dataset is deleted at the end of the JCL'
You are of course quite correct, I answered before thinking it through. |
|
Back to top |
|
|
mucka
New User
Joined: 11 Apr 2017 Posts: 5 Location: USA
|
|
|
|
Could you just keep one version of the dataset and get your information and then delete it manually? |
|
Back to top |
|
|
ashek15
New User
Joined: 20 Oct 2015 Posts: 48 Location: USA
|
|
|
|
Actually these are many prod jobs for which we need the allocation details and hence it's not feasible to change all of them just for getting dataset allocation details and then again change them to their current state. Moreover, we don't really want to keep these datasets occupying space on dasd as these are temporary datasets and should be deleted as soon as jcl execution is completed. Hence looking for other feasible options. Thanks. |
|
Back to top |
|
|
Nic Clouston
Global Moderator
Joined: 10 May 2007 Posts: 2455 Location: Hampshire, UK
|
|
|
|
Why not temporarily add an IDCAMS step at the end of the job (before data set deletion) to list all the data sets and their properties, used in the job? |
|
Back to top |
|
|
Robert Sample
Global Moderator
Joined: 06 Jun 2008 Posts: 8696 Location: Dubuque, Iowa, USA
|
|
|
|
Quote: |
Hence looking for other feasible options. |
The usual method these days is to create a storage pool for work data sets. Allocate an appropriate number of volumes and let the system manage the space. How many x37 abends a week (or month) does your site have that there is a need to do what you're doing?
The basic problem you have is that for temporary data sets, there is NO easy way on z/OS to tell how much disk space they use while the job is executing, and when the data set is deleted when the job is complete you cannot go back and find out the space needed. Your storage management site support group should be involved, and they can size the storage pool based on all the applications running on the system. |
|
Back to top |
|
|
Willy Jensen
Active Member
Joined: 01 Sep 2015 Posts: 712 Location: Denmark
|
|
|
|
If you really dont know the initial size and your goal is to save temporary space, then consider changing the allocation to say from (CYL,(200,300)) to (CYL,(100,400)), so that the initial allocation is smaller. And going back to SMF type 15, if you can find the count of blocks written then you might be able to calculate the space used. |
|
Back to top |
|
|
David Robinson
Active User
Joined: 21 Dec 2011 Posts: 199 Location: UK
|
|
|
|
SMF records will show you how many extents the dataset has used. From that it's easy to assess whether you are likely to be running out of space any time soon.
That's the approach I have used in a couple of different sites. Pick a percentage, say 50% or 75%, and if 's using more than that % of the maximum possible extents it could have, raise an alert.
Note that the SMF record will also show you how many volumes it can extend across. So, if the volume count is 5, you know that the dataset could have 80 extents, for example. |
|
Back to top |
|
|
vasanthz
Global Moderator
Joined: 28 Aug 2007 Posts: 1742 Location: Tirupur, India
|
|
|
|
Hi David what smf record types? |
|
Back to top |
|
|
steve-myers
Active Member
Joined: 30 Nov 2013 Posts: 917 Location: The Universe
|
|
|
|
vasanthz - The only SMF record type we've been discussing here is the type 15 record.
There is one field of possible interest, though I don't know how reliable it is: SMF14NTU. Don't worry about the 14 in the label; type 14 and 15 are basically identical so type 15 records are mapped by type 14 symbols. I use types 14 and 15 in some of my programs, but I have never used SMF14NTU. |
|
Back to top |
|
|
Rohit Umarjikar
Global Moderator
Joined: 21 Sep 2010 Posts: 3049 Location: NYC,USA
|
|
|
|
Quote: |
we need this info to ensure we can track/monitor the sequential dataset usage and proactively allocate the appropriate space so that any future job failures can be avoided due to B37, D37 etc. |
1.You can give maximum size in secondary allocations based on the stats of your earlier failures and use MXIG and RLSE to release unused space.
2.However, You also may need to analyze the size of the input dataset and allocate same or more to your work data.
3.DATACLAS are site specific, so you need to speak to your storage people if you could use that and then don't worry about 1,2 or anything for that matter. |
|
Back to top |
|
|
|