Re: Save while active on 6.1

[CRPence <CRPbottle@xxxxxxxxx>]

rob@xxxxxxxxx wrote:
> If you are running save while active on 6.1 get this ptf and get
> it installed
>   6.1.1 - MF49320
>   6.1.0 - MF49321
>
> They are immediate with no special action. Otherwise a save while
> active operation may hang on some object. If it hangs then you
> will not be able to use save while active on that object until
> your next ipl. The ptf will prevent the hang.
>
> Night 1: Backup hung on the object.  Canceled backup.
> Night 2: Backup complained that the object was unable to be saved
> using save while active. CPF3795, Error summary code C4E6.
> Object is not really damaged. Just some save while active flag
> associated with the object. Regular save works fine.

  An object having a flag indicating that SWA is active on the 
object when in fact SWA is not active against that object is 
referred to as "logical damage"; i.e. by the given description, the 
object was in an inconsistent\illogical state, and therefore the 
object was indeed "damaged".  The term "damage" is utilized for more 
than the types [full & partial] which prevent SAVxxx.  Depending on 
the specific /logical damage/ as defined by the component which owns 
an object type or a function, saving the object may or may not be 
prevented.  In the case of an object tracked under SWA, the non-SWA 
save is allowed because the actual object and its data should be 
valid\intact even if the status of its changed page tracking under 
the SWA are indeterminate, due to an abnormal failure of the prior 
save-while-active request which placed the object under changed-page 
tracking.  The LD [Load\Dump] component deems that the object 
already under SWA can not be saved under SWA, but that a non-SWA 
dump [i.e. save] action can be accommodated given the design of SWA; 
i.e. just dump the object and its data segments per normal, because 
only a SWA request must redirect to the sync pages which are "called 
into question" because of the previous failing\canceled SWA request.

  So apparently there was, as I had questioned as possibly if, 
there was a defect:
http://archive.midrange.com/midrange-l/201003/msg00062.html re:
http://archive.midrange.com/midrange-l/201002/msg01490.html

  FWiW the noted PTFs are from the APAR MA39127 which is on the PSP 
HIPer list for the error msgcpf3795 ecc4e6 [or rcc4e6 or c4e6] as a 
symptom diagnosing the side-effect of a prior interrupted SWA; in 
this case, from a prior canceled HANG [or WAIT] condition.

  Links to the HIPer APAR & a PSP document showing the HIPer 
designation text:
http://www-01.ibm.com/support/docview.wss?uid=nas255954f589e98fc33862576d70042481d
http://www-912.ibm.com/s_dir/sline003.NSF/7f02fa4a8460c21f86256c450052b184/288a4b2131a57da4862573df00528536?OpenDocument&Highlight=2,SF99610

Regards, Chuck