This is the 15th semiannual presentation of Bob Shannon's and Carl Youngren's Bit Bucket feature. In it they list in rapid fire a bunch of oddities, things of interest, and stuff that didn't seem to fit anywhere else in SHARE. They are both long time SHARE contributors, and for this and other efforts they were awarded the SHARE President's Award at this meeting.

Without further ado, the Bit Bucket:

• Try extended indirect volume serials! They are useful for SYSRES overflow volumes, and maybe for DLIB overflow volumes as well.
• There is an undocumented "D LLA" command available in release 2.4. Almost every new function coming out of IBM these days has a corresponding "DISPLAY" command that is associated with it. So if you don't see any DISPLAY documentation for a new function, sometimes you can just try it and see what you get.
• The IODF suffix doesn't have to be in the hex range 00-BF. It can have some magic values with magic meanings: "**", "--", "++" and "==" all do different things. This is certainly weird, but it may be useful.
• The Catalog Search API program product was integrated into DFSMS 1.4. Nobody was buying it.
• OS/390 1.3 changed the behavior of system dumps. Programs running in keys 0-7 don't get dumped anymore. This is not documented, but adversely affects your ability to debug third party program products. Two new RACF profiles protect the ability to dump key 0-7 address spaces. If you feel that this has worked for 20 years, and isn't broken now, then you can simply allow the dump to take place by specifying:

     IEAABD.DUMPAKEY UACC(READ)
     IEAABD.DUMPAUTH UACC(READ)
  

• PTF UW37471 will introduce problems with block counts on tape. Put on APAR OW28282! Now! This is a high impact problem.
• Recent maintenance on IEBGENER causes it to drop records, with no obvious side effects. YOW! How could IBM let this happen??!
• People running CA's TMS product are facing a "year 98" crisis. TMS uses a variety of 98.xxx experation dates for special processing. Late this year a bunch of tape libraries are going to explode. Forewarned is forearmed...
• On 9672 G3 and G4 CPUs, you can get access to the hardware console through a browser! This has to be enabled by your CE through a microcode update called "driver '98". There isn't any really fancy security implemented in this; passwords are required, but they are transmitted through the network in the clear. So you should probably restrict access to a local LAN that you have control over - unless you want to give some kid with a little too much time on his hands the ability to power your system off.
• IBM has extended currency for all OS/390 releases through January 31, 2001. They know that most installations will be dealing with a date crisis, and that they don't have time to be chasing operating system upgrades.
• Historically, installations IPL their MVS systems using "LOAD" rather than "LOAD/CLEAR". No sense clearing memory when the first thing that MVS NIP does is clear it again, right? Well now it seems that there is a reason to use LOAD/CLEAR: some installations have reported a memory leak in HSA, which eventually becomes bad enough to refuse IPL. The fix is to use LOAD/CLEAR, which causes the PR/SM hypervisor to clean up the HSA. Weird.
• The Great Load Library Blocksize Debate continues. The most recent recommendation is 32760. We'll see how long this one holds out.