This problem is caused by a few things that came together in this one job...
- The allocation amounts were quite small.
- EOV_VSAM was invoked frequently.
- A large number of volumes were defined as candidates.
We have some jobs REPROing large VSAM files that when running under VAM are taking much longer during EOV processing.
Same jobs running where VAM is bypassed run much faster.
When Vkgparm PLSOPT24 (Y) was specified, the job runtime was equivalent to when VAM was bypassed.
N (default) - Launch EOV_VSAM even if candidate is available.
Y - Prevent launching EOV_VSAM when a candidate volume is available.