News
Abstract
Some workloads* were experiencing poor performance after loading TR12 (V7R4) or TR7 (V7R5). Hiper fixes were created to resolve the problem and the TR distribution fix was marked defective. If partition is at affected TR level and a performance issue has been observed, plan an IPL to apply SJ09037 (V7R4) or SJ09038 (V7R5).
*Examples of workload include, but not limited to, SQE/SQL queries, RUNQRY, jobs running in pools with an expert cache setting of *CALC.
Content
How can I determine if my partition is at the affected TR level?
All partitions with MJ99312 or MJ99407 applied prior to March 17, 2026 are affected. It is possible the TR was applied if a group PTF was recently applied. It is best to verify using the following CL commands:
V7R4: DSPPTF LICPGM(5770999) SELECT(MJ99312)
V7R5: DSPPTF LICPGM(5770999) SELECT(MJ99407)
17 Mar 2026 DEFECTIVE PTF: SJ05258 for IBM i 7.4
LICENSED PROGRAM = 5770SS1
APAR NUMBER: DT465567
USERS AFFECTED: All IBM i operating system users.
REASON DEFECTIVE: After applying the PTF, systems running certain workloads may experience performance degradation, including excessive paging in the main storage pools, or a system hang.
RECOMMENDATION: Apply fixing PTF SJ09037.
Note that PTF SJ09037 is a distribution requisite of the latest Technology Refresh (TR) PTF. If you have applied the latest TR PTF to your system, you should apply the fixing PTF SJ09037 for IBM i 7.4.
17 Mar 2026 DEFECTIVE PTF: SJ07685 for IBM i 7.5
LICENSED PROGRAM = 5770SS1
APAR NUMBER: DT465567
USERS AFFECTED: All IBM i operating system users.
REASON DEFECTIVE: After applying the PTF, systems running certain workloads may experience performance degradation, including excessive paging in the main storage pools, or a system hang.
RECOMMENDATION: Apply fixing PTF SJ09038.
Note that PTF SJ09038 is a distribution requisite of the latest Technology Refresh (TR) PTF. If you have applied the latest TR PTF to your system, you should apply the fixing PTF SJ09038 for IBM i 7.5.
What hiper fixes are part of the new distribution fixes?
The superseding distribution PTF includes all of the fixes for the defective issue.
| V7R5M0 | PRE 5770999 | MJ08160 | Applies to pools using expert cache setting of CALC, more IO pending faults for the same number of async DB reads. |
| V7R5M0 | PRE 5770999 | MJ08470 | Large SQE sorts in a memory pool, *CALC or *FIXED, will take longer. |
| V7R5M0 | PRE 5770999 | MJ08929 | Temp pages are not being written out of memory fast enough to satisfy demand for same. RUNQRY performance is significantly impacted as well as the performance of other non-RUNQRY functions. |
| V7R5M0 | PRE 5770999 | MJ08076 | After DPO of an active LPAR, the affinity fixup code can get into a looping state checking for a page state, and can hold storage management gates causing some memory related functions to hang. |
| V7R5M0 | PRE 5770999 | MJ07682 | Load source NVMe FW update fails with B6005121. |
| V7R4M0 | PRE 5770999 | MJ08158 | Applies to pools using expert cache setting of CALC, more IO pending faults for the same number of async DB reads. |
| V7R4M0 | PRE 5770999 | MJ08469 | Large SQE sorts in a memory pool, *CALC or *FIXED, will take longer. |
| V7R4M0 | PRE 5770999 | MJ08930 | Temp pages are not being written out of memory fast enough to satisfy demand for same. RUNQRY performance is significantly impacted as well as the performance of other non-RUNQRY functions. |
| V7R4M0 | PRE 5770999 | MJ08077 | After DPO of an active LPAR, the affinity fixup code can get into a looping state checking for a page state, and can hold storage management gates causing some memory related functions to hang. |
| V7R4M0 | PRE 5770999 | MJ08284 | Contents of memory pages are removed, stolen, too soon. Not dependent on memory paging pool setting (*CALC or *FIXED). Any number of performance symptoms, from SQE performance to native I/O issues. Especially in scenarios where the OS pre-brings data into memory, those pages will be stolen by the time they are needed. |
Do the fixes require an IPL?
Yes
If I am going to IPL, are there other hiper performance fixes to consider applying?
| V7R4M0 | 5770999 | MJ08541 | Applies to SQE table scans. Pages, in the pool, are being forced out which leads to page faulting. DT460082 |
| V7R5M0 | 5770999 | MJ08539 | Applies to SQE table scans. Pages, in the pool, are being forced out which leads to page faulting. DT460082 |
| V7R5M0 | 5770999 | MJ08635 | The "reprime" logic in the DspScan node was not re-bringing the proper range when page faults are detected. IMPORTANT NOTE: if this fix is applied immediately, the fix will only be applied to newly created access plans; all access plans are cleared and newly created at IPL time. DT461323 |
How soon should an IPL be scheduled to apply the fix?
If the workloads are stable with no performance concerns, you can wait for a convenient IPL window.
If the workload is expected to increase or performance is impacted, expedite applying the fix.
Should I apply the fixing PTFs if I am not at TR level 12 or 7?
No
Will IBM assess our system and tell us if we are exposed to this issue?
No. If the latest TR is applied, the defect is in play. The client will need to decide if the performance impact is poor enough to warrant expediting the IPL to apply the fixes. The client will need to choose how to limit their exposure.
What should I do if I have more questions?
If the fixing PTF did not resolve the performance problem, your workload was not affected by the performance issue described in this document. If you need assistance identifying a performance issue, please open a case with the IBM i Global Support Center. As noted previously in this document, the iGSC will not perform an assessment or compare pre/post data. Instead, our support teams will guide you to investigate if a performance problem is happening and which proactive steps are best to implement in your environment.
Related Information
Was this topic helpful?
Document Information
Modified date:
24 March 2026
UID
ibm17267270