An SDC2 abend on a JES2 warm start? Sounds very familiar. I spent quite a bit of time with support over that. Here is what was finally determined:
OA58190 has changed the JES2 checkpoint structure, which is independent of user activity. Such internal changes are necessary in order to support new functionality. Therefore all JES2 members accessing such a checkpoint thereafter need the mentioned compatibility APAR OA59665/UJ03635. Thus..... If these APARs/PTFs are applied, there is no falling back. And a fallback attempt causes the SDC2 abend. Matthew On Sun, 28 Feb 2021 08:08:55 -0800, Ed Jaffe <[email protected]> wrote: >On 2/27/2021 7:48 PM, Mark Jacobs wrote: >> >From the FM; >> >> The summary dump data contains data captured closest to the time of the >> failure. If you reply Y to use this data, IPCS will not be able to display >> storage keys using the DISPLAY(MACHINE) parameter. > >Yeah, but... > >I don't experience issues displaying keys using DISPLAY(MACHINE) when >responding 'Y' to BLS18160D. > >I tried it just now. I initialized a dump still lying around from >yesterday (SLIP to capture an SDC2 abend in an SRB) and responded 'Y' to >the BLS18160D prompt: > > IKJ56650I TIME-07:42:42 AM. CPU-00:00:03 SERVICE-22127 >SESSION-00:00:47 FEBRUARY 28,2021 > BLS18122I Initialization in progress for >DSNAME('SYS3.DUMP.D210227.T222342.EDJX2.S00008') > BLS18124I TITLE=SLIP DUMP ID=EEJ6 > BLS18223I Dump written by z/OS 02.04.00-0 SLIP - level same as IPCS level > BLS18558I This redactable dump has not been post-processed to protect >sensitive data > BLS18222I z/Architecture mode system > BLS18160D May summary dump data be used by dump access? Enter Y to >use, N to bypass. >y > BLS18255I Dump Init Elapsed Time CPU Time > Input I/O 00:00:01.205379 00:00:01.041953 > DDIR 00:00:00.842453 00:00:00.605543 > BLS18123I 161,077 blocks, 670,080,320 bytes, in >DSNAME('SYS3.DUMP.D210227.T222342.EDJX2.S00008') > IKJ56650I TIME-07:44:17 AM. CPU-00:00:07 SERVICE-42085 >SESSION-00:02:21 FEBRUARY 28,2021 > BLS18224I Dump of z/OS 02.04.00-0 - level same as IPCS level > *** > >STATUS shows PSW and registers for the DC2 abend: > >CPU STATUS: >PSW=47045000 80000000 00000000 019A7318 > (Running in AR, key 0, AMODE 31, DAT ON, SUPERVISOR STATE) > Enabled for PER I/O EXT MCH > ASID(X'0072') 019A7318. IEANUC01.IAXV6+24D0 IN READ ONLY NUCLEUS > ASCB114 at FBD400, JOB(EDJX2), for the home ASID > ASXB114 at AFD000 for the home ASID. No block is dispatched > HOME ASID: 0072 PRIMARY ASID: 0072 SECONDARY ASID: 0072 > > General purpose register values > Left halves of all registers contain zeros > 0-3 84000000 84DC2000 00000001 00000000 > 4-7 00000000 029B9000 00FBD400 01DDDF00 > 8-11 00000000 042FCC30 00000000 08004000 > 12-15 019A8550 042FE178 042FE000 66004020 > >SUMDUMP captures storage on both sides of the PSW and every register. I >have 'IP SETDEF DISPLAY(MACHINE)' in effect. Listing the PSW via 'IP L >019A7318' returns KEY(08): > >LIST 019A7318. ASID(X'0072') LENGTH(X'2000') AREA >ASID(X'0072') ADDRESS(019A7318.) KEY(00) >019A7318. B24D001C E31D0000 00045811 00889110 |.(..T........hj.| >019A7328. 1008A784 0009C019 04C0D000 C0F9FE00 |..xd..{..{}.{9..| > >'IP L 5R!' returns KEY(00): > >LIST 029B9000. ASID(X'0072') LENGTH(X'2000') AREA >ASID(X'0072') ADDRESS(029B9000.) KEY(08) >029B9000. C9C1E7C3 D7E3C2D3 00000012 00000000 |IAXCPTBL........| >029B9010. 00000012 00000040 042FC000 00000000 |....... ..{.....| > >'IP L 7R!' returns KEY(00): > >LIST 01DDDF00. ASID(X'0072') LENGTH(X'2000') AREA >ASID(X'0072') ADDRESS(01DDDF00.) KEY(00) >01DDDF00. D9C3C540 002E65A3 00000FFE 00000EB6 |RCE ...t........| >01DDDF10. 0000050B 00350202 0000142E 000028C0 |...............{| > >'IP L 9R!' returns KEY(00): > >LIST 042FCC30. ASID(X'0072') LENGTH(X'2000') AREA >ASID(X'0072') ADDRESS(042FCC30.) KEY(00) >042FCC30. 00000000 009A0000 00000000 00000000 |................| >042FCC40. 00000000 042FC000 00000000 042FC004 |......{.......{.| > >Same with all of the other registers that look like addresses. Not one >does not show the KEY information. I get similar results when processing >dumps in TCB mode. Pretty much everything pointed to by the RTM2WA is >included in the SUMDUMP and KEY information seems to be routinely provided. > >Perhaps there is some pathological situation where that information >doesn't display? > >-- ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
