Page 1 of 1

IAM file running out of space...

PostPosted: Fri Jul 23, 2010 5:31 pm
by Bad Man
Hi How do we know before hand that an PS and IAM file will run out of spcae i.e. get a B37...Please tell me!!

Re: IAM file running out of space...

PostPosted: Fri Jul 23, 2010 5:37 pm
by Robert Sample
I can't talk to IAM files -- I've never used them.

In general, it is not possible to determine in advance if a sequential file is going to run out of space. If you monitor the file over time and see it adding an extent every few days (or weeks), then you should delete it and give it more space before it hits 16 extents. However, a file may run perfectly fine for months and then get a space condition because someone attempted to load 100 or 1000 times as much data as normal into the file. Hence you cannot always predict in advance which file(s) will get space abends.

Re: IAM file running out of space...

PostPosted: Sat Jul 24, 2010 12:13 am
by steve-myers
Like Mr. Sample, I've never used IAM. IAM's documentation may provide methods to estimate dataset requirements. I can almost guarantee it will not be less than the sequential file you're using as its base!

Re: IAM file running out of space...

PostPosted: Sat Jul 24, 2010 1:12 pm
by expat
IAM is the Innovation Data Processing version of VSAM, and like good old traditional VSAM can be queried using IDCAMS LISTCAT

Re: IAM file running out of space...

PostPosted: Sun Jul 25, 2010 1:22 am
by Bad Man
Mr. Sample I think I will have to read more as I found a code in FDRABR to report on PS running out of space by 10% and that leaves me more wanderingI will have to read on how that report or what it can do....

Re: IAM file running out of space...

PostPosted: Sun Jul 25, 2010 3:28 am
by dick scherrer
Hello,

I suspect that the report from FDRABR will tell you when an allocated file is 90% full - which is probably not what you want. . . This would have been handy in the "old days" when an allocation was an allocation and dynamic extension did not happen. . .

Also, as has already been mentioned, no tool will be able to predict spikes in volume. . .

If your dasd is not well managed, you may experience x37 abends due to some other processes taking the space that might have been used for your process. Indeed, even if there appears to be enough space when you "measure", something else may have taken it before your process can get to it.

On a different note, most times a sequential file is written, it is a newly created output file - not some existing file (yes MOD can be used to append, but i only use MOD intra job and not for datasets that are already "there"). If a new dataset is being written, allocating a moderate primary and overly large secondaries with RLSE should prevent the x37 and free any unneeded space when the file is closed.

Re: IAM file running out of space...

PostPosted: Mon Jul 26, 2010 2:44 am
by cssaus
As other posters have stated, it's not possbile to accurately predict when any file is going to incur an x37 space abend. However it is possible to minimize the likelihood of an abend occuring using the facilities of DFSMS or even tools such as STOPX/37 etc.

IAM has a feature called "DYNAMIC SECONDARY SPACE ADJUSTMENT" which will dynamically adjust the secondary space quantity for Enhanced Format files. This is documented on page 02-11 of the IAM Version 8.1 User Reference Manual.