Paul, That is a very illuminating table! I had not seen that before and it certainly appears to show that the 257 is about right. There were a few other allocations in the job step of course.
Thank you Paul. The collective intelligence of this group is a valuable resource. Lennie -----Original Message----- From: IBM Mainframe Discussion List <[email protected]> On Behalf Of Paul Feller Sent: 10 December 2024 21:02 To: [email protected] Subject: Re: DYNAMNBR - parallel Well, the 257 datasets (DDs) seems about right if they all are using a volume count of 59. Based on the volume count you would be allowed a maximum of 259 DDs in the step. That's based on a TIOT size of 64. "So while I cannot reconcile the numbers of allocated datasets with the number of TIOT slots (257 x 59 = 15163 and 15163 > 3273) I think the DVC is likely to be the cause of the problem." I hope my past of the table from the MVS Initialization and Tuning Reference for z/OS 3.1 comes through. Table 14. Relationship size of TIOT and maximum number of DDs allowed Dec Hex Size of TIOT Maximum number of single units DD allowed Maximum number of DDs allowed when every DD requests the maximum number of units (59) 48 30 49152 (48 K) 2454 194 56 38 57344 (56 K) 2864 227 64 40 65536 (64 K) 3273 259 Based on the information you have provided I would agree that the volume count is causing an issue for you. I find it interesting that SMS could not get in to alter the DATACLAS of the files during allocation time. But I'm not an expert in SMS. Paul -----Original Message----- From: IBM Mainframe Discussion List <[email protected]> On Behalf Of Lennie Bradshaw Sent: Tuesday, December 10, 2024 8:45 AM To: [email protected] Subject: Re: DYNAMNBR - parallel Hi Martin, These data sets are allocated by zSecure, so I would have to enquire of them whether they attempt to multitask their allocations. My rather old history tells me that allocation used to serialise quite substantially, thus making any subtask allocation process wait, but maybe things are different now. Lennie -----Original Message----- From: IBM Mainframe Discussion List <[email protected] <mailto:[email protected]> > On Behalf Of Martin Packer Sent: 10 December 2024 13:50 To: [email protected] <mailto:[email protected]> Subject: Re: DYNAMNBR As a matter of interest, Lennie, are these data sets allocated and de-allocated sequentially, or all at the same time? Question to others: Does that matter in terms of (X)TIOT slots? Cheers, Martin From: IBM Mainframe Discussion List <[email protected] <mailto:[email protected]> > on behalf of Lennie Bradshaw <[email protected] <mailto:[email protected]> > Date: Tuesday, 10 December 2024 at 13:45 To: [email protected] <mailto:[email protected]> <[email protected] <mailto:[email protected]> > Subject: [EXTERNAL] Re: DYNAMNBR Scott, Thanks again for the detail. Firstly a couple of questions. 1. Can XTIOT only be requested by using Dynamic Allocation? 2. Is XTIOT only available to authorised callers? Looking at the allocation failures in question I have, IEFA111I JJJJJJJJ IS USING THE FOLLOWING JOB RELATED SETTINGS: SWA=ABOVE, TIOT SIZE=64K,DSENQSHR=DISALLOW,GDGBIAS=JOB At the failure I have, IKJ56220I DATA SET -----REDACTED----- NOT ALLOCATED, TOO MANY DATA SETS+ IKJ56220I MAXIMUM NUMBER OF DATA SET ALLOCATIONS ALLOWED BY YOUR SESSION HAS BEEN REACHED, YOU SHOULD FREE UNUSED DATA SETS. This was a batch job requesting allocation of 312 similar data sets. The count of IGD103I messages indicates 257 were allocated. Crucially I see all these data sets were allocated with a DATACLAS value which has a DVC of 59. So while I cannot reconcile the numbers of allocated datasets with the number of TIOT slots (257 x 59 = 15163 and 15163 > 3273) I think the DVC is likely to be the cause of the problem. Looks like I have no way of changing the DATACLAS used for those allocations; It cannot be overridden nor can it be changed. So it looks like the only way to handle these datasets is to copy them using a different DATACLAS. Does that look correct to you? Lennie -----Original Message----- From: IBM Mainframe Discussion List <[email protected] <mailto:[email protected]> > On Behalf Of Scott Ballentine Sent: 09 December 2024 14:05 To: [email protected] <mailto:[email protected]> Subject: Re: DYNAMNBR I'll throw a few more random things out here that might be helpful. S99ERROR code x'043C' indicates that the DYNAMNBR value is being exceeded. If you're getting a different error... then it's probably not worth digging around this value. Allocation requests that ask for an XTIOT don't take space from the 64K (or smaller) TIOT. Also it's a good idea to request DSABs above-the-line and actual UCBs, otherwise you end up with a lot of control blocks in 24-bit storage. There are a few ways to do this, but the easiest is to set the S99DXACU flag in the S99RB - that gives you all three of those options. Also make sure the job class has SWA=ABOVE. Take a look at the number of data sets that are able to be allocated at one time. You can make some guesses about what limit you're running into based on that. The TIOT can hold a maximum of 3273 DDs (and less if they are multi-volume data sets or a DVC is in use.) If they're getting past that limit (maybe into the tens of thousands of allocations), it's more likely that these are XTIOT requests, but some of the other control blocks (such as the DSABs) are going into 24-bit storage. -Scott Ballentine IBM z/OS Device Allocation [email protected] <mailto:[email protected]> ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] <mailto:[email protected]> with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] <mailto:[email protected]> with the message: INFO IBM-MAIN Unless otherwise stated above: IBM United Kingdom Limited Registered in England and Wales with number 741598 Registered office: Building C, IBM Hursley Office, Hursley Park Road, Winchester, Hampshire SO21 2JN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] <mailto:[email protected]> with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] <mailto:[email protected]> with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
