SNAP runs on all BS2000 servers.
The data saved thus far is preserved and the fragmented SNAP dump is made available to the diagnostic staff in the following cases
the data to be saved by SNAP (maximum 1 GB) cannot be written within the time set for SNAPTIME (Default: 24 seconds)
the SNAPFILE file is too small
an internal error occurs
In this case, message NSP1010 is output. For serious errors in SNAP, the SNAP process is aborted.
Within BS2000 there are certain facilities which recognize if the BS2000 system has been inoperable for longer than a predefined time. SNAP deactivates BS2000 so that it can secure diagnostic data under its own runtime control. However, this does not been that BS2000 is “dead”. When defining intervals for the detection of system failures, it should be borne in mind that SNAP-EXEC is still active (as a “stand-in”) even though BS2000 is inactive.
In this context, special attention is drawn to the product HIPLEX MSCF, which can be used for vital-sign monitoring for BS2000 systems in conjunction with the MSCF configuration parameter FAIL-DETECTION-LIMIT.
BS2000 OSD/BC and SNAP support a working memory of more than 2 GB.
As a result, not only the class 1 and class 3 memory may be contained in the SNAP dump, but also the entire resident class 4 memory if the time limit set in the system parameter SNAPTIME is not exceeded and the SNAPFILE file is sufficiently large.
The system parameter SNAPTIME is limited to the interval of 8 to 24 seconds. When too low a value is set, it is rounded up to 8 seconds, and when too high a value is set, it is rounded down to 24.