| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> RE: virtual circuit wait/shared server
Barb,
Common complaint that I've heard with MTS is that when a user asks for = something if there is no shared server in a wait(common) state, then = they will have to wait for one. It sounds to me like at your peak time = there are insufficient servers to process all of the requests in a = "timely" manner. Some time ago I modified my friendly Oracheck program = to watch the shared servers every time I comes around. It revealed that = starting & killing off shared servers is one of those tasks that Oracle = does not have a high priority on. Namely at peak load, when it should = spawn 3 or 4 shared servers it won't for an appreciable time. Seems = like the kernel is taking a wait & see approach to the matter. My fix = was to see is a specified minimum number of shared servers were in = wait(common). If not adjust the min_shared_servers parameter with the = alter database command. Worked like a charm. Here's an experiment you = can try to see if that's what's bothering you. At the start of the = morning alter the database to have min_shared_servers =3D 2 times the = initial value. Then at noon reduce them to normal. Should kill your = complaints.
Dick Goulet
Senior Oracle DBA
Oracle Certified 8i DBA
 |
 |