If the SDM read process stalls, or if the application update rate continually exceeds the capability of the SDM to offload application updates, the amount of cache consumed in the subsystem will continue to grow.
If the SDM continues to read data, except at a relatively slow rate, the dead system data mover timer, discussed in Application protection, will not be invoked because the SDM is still running, but not keeping pace with the cache consumption rate.
Over time, the residuals will grow for all primary volumes and a long busy condition may be reached. This condition can result from excessive cache consumption (on an IBM® ESS, copy service functions are limited to 60% of the cache resource) or an excessive number of records in cache for a particular scsession (on the IBM ESS, the limit is 63K records).
It is possible for the primary disk subsystem to remain in a long-busy condition for 80% of the TIMEOUT value for a given session. When this occurs, the SDM suspends the volumes in the session. If this action does not correct the condition, the session ends for non-ESS subsystems and is suspended for ESS subsystems.