library cache latches on RAC
Date: Thu, 22 Jan 2009 12:01:24 -0600
Message-ID: <BAY103-W14345A5EFA13C1B09B0BFBF8CE0_at_phx.gbl>
We have a newly implemented RAC environment on a primary revenue system. We have a reservation center using Oracle forms and we have external requests coming in to the environment from the internet for reservation availability. There are 5 nodes on the RAC cluster. Two nodes run services for the users in the reservation center and two nodes run services for the external in-coming requests. We are seeing sporadic library cache latch waits on the nodes for the external queries. However, when one of those nodes experiences latch waits, the users on the reservation floor (who are on seperate nodes) also experience "clocking" or extended waits.
If a user was on the node experiencing latch waits, I would expect that user to experience slight delays. However, I don't undertand why users on the other nodes are also experiencing delays. For example, if node 3 experience latch waits, then users on node 1 and node 2 also appear to be impacted. Can someone help me understand how and why?
Second question. The latch waits are caused by the queries coming in from internet sources. They occur sporadically throughtout the day and only last 2-4 minutes. I believe the sessions are created and terminated for each in-coming request. Any ideas how I can diagnose the cause of the latch waits in such a short period. Tracking all incoming sessions is impractical because of the quantity of the sessions.
Any insights you have will be GREATLY appreciated.
Regards,
Nancy Iles
Omni Hotels
Windows Live™: E-mail. Chat. Share. Get more ways to connect. http://windowslive.com/explore?ocid=TXT_TAGLM_WL_t2_allup_explore_012009-- http://www.freelists.org/webpage/oracle-l Received on Thu Jan 22 2009 - 12:01:24 CST