Joined: 06 Jun 2008 Posts: 8214 Location: Dubuque, Iowa, USA
Short answer (which is always true for S-O-S problems): figure out what's eating up the storage and change the program(s) to resolve it -- or add more memory to the region.
Run a STAT in the region and find out which storage pool is having the problem. If it's one of the above-the-line pools like ECSA, and you can increase the memory to that pool, that could be a quick fix. If it's one of the below-the-line pools, and the problem just recently started, you need to look at application program changes recently and identify which program(s) are using up the storage. Getting the application to run above the line is often the best way to resolve below-the-line S-O-S issues.
If someone recently made a region change (say increasing max tasks from 50 to 250, for example), without changing the memory allocation for the region appropriately, that could also cause S-O-S issues.
So it could be an application OR a system issue -- without doing investigation on site, you cannot really say which, nor how to fix it. But you definitely don't want S-O-S situations to continue as they have big impacts on response times.