As you tune the RAPCache server by adding named caches or increasing the sizes of existing caches, the server may stop rebooting at some point. The memory allocated to the server is large enough to accommodate the additional cache size, and the operating system level parameter SHMMAX (the maximum size in bytes for a shared memory segment) is also large enough, but the server does not restart. The problem may be that the operating system level parameter SHMALL, the system-wide maximum of shared memory pages, is not large enough. To resolve this issue, increase the value of SHMALL in the file /proc/sys/kernel/shmall. Modifying /proc/sys/kernel/shmall requires root permission.
