Kbase 20851: Considerations with Use of Multiple Semaphore Sets
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
13/05/2005 |
|
Status: Unverified
GOAL:
Some considerations when using multiple semaphore sets.
FIX:
Some of the things to consider are:
- Without multiple semaphore sets, Progress uses one semaphore set for each database. The kernel parameters must be set to allow a semaphore ID per database. The maximum semaphore set size must be large enough to accommodate the number of users.
- When using multiple semaphore sets, divide the single large set into several smaller ones. Remember to allow for more semaphore IDs, but the maximum semaphore set size can be reduced.
- The kernel data structures for semaphores are fairly small, so there should be no problem to make them larger than the exact amounts you need.
- The trend in UNIX is to make most kernel settings dynamic (semaphores and shared memory included), so it is not necessary to set them at all. Solaris and AIX are examples.