IFS Node hanging (Needs Database restart)

Question

Hi ,
We have system which was working pretty good.Recently we started getting these problems with the iFS node. We are not able to connect to iFS and our client applications are getting the following error.
We have the follwing configuration.
1) Oracle iFS 9.0.1
2) Oracle DB 9.0.1
3) We use iFS SDK as middle tier and run iFS service on middle tier
4) We have some custom agents running on serverside (Node).
5) Node runs with default service which uses "SmallServiceConfiguration".
6) Clients also use service property file which has "SmallServiceConfiguration".
Here is the exception which we are getting.
java.lang.Exception: Exception while trying to get Connection
oracle.ifs.common.IfsException: IFS-21008: Unable to connect to iFS service
oracle.ifs.common.IfsException: IFS-10151: Unable to perform authentication
oracle.ifs.common.IfsException: IFS-10621: Unable to acquire library connection
oracle.ifs.common.IfsException: IFS-10630: No unallocated connections (timed out waiting)
     at oracle.ifs.server.ConnectionPool.acquireLibraryConnection(ConnectionPool.java:580)
     at oracle.ifs.server.IfsCredentialManager.authenticate(IfsCredentialManager.java:237)
     at oracle.ifs.server.DirectoryService.authenticate(DirectoryService.java:527)
     at oracle.ifs.server.S_LibraryService.connect(S_LibraryService.java:2495)
     at oracle.ifs.beans.LibraryService.connect(LibraryService.java:977)
     at com.cisco.as.ifs.util.CiscoService.connect(CiscoService.java:79)
Looking forward for your reply.

Answer 1

Hi Luis,
We have web applications which creates library session for each user who logs in. The library session gets disconnected when users HttpSession gets invalidated.
We did check how many connections were open from the application by running query against V$session table and there were around 30 inactive connections.
Rightnow here are the details of properties which are set.
# Service data cache
IFS.SERVICE.DATACACHE.Size=7500
IFS.SERVICE.DATACACHE.NormalTrigger=5000
IFS.SERVICE.DATACACHE.UrgentTrigger=5500
IFS.SERVICE.DATACACHE.EmergencyTrigger=6000
IFS.SERVICE.DATACACHE.PurgeTarget=4000
# Writeable connection pool
IFS.SERVICE.CONNECTIONPOOL.WRITEABLE.MinimumSize=2
IFS.SERVICE.CONNECTIONPOOL.WRITEABLE.TargetSize=5
IFS.SERVICE.CONNECTIONPOOL.WRITEABLE.MaximumSize=10
IFS.SERVICE.CONNECTIONPOOL.WRITEABLE.TargetSizeTimeout=1000
IFS.SERVICE.CONNECTIONPOOL.WRITEABLE.MaximumSizeTimeout=10000
# Readonly connection pool
IFS.SERVICE.CONNECTIONPOOL.READONLY.MinimumSize=2
IFS.SERVICE.CONNECTIONPOOL.READONLY.TargetSize=5
IFS.SERVICE.CONNECTIONPOOL.READONLY.MaximumSize=10
IFS.SERVICE.CONNECTIONPOOL.READONLY.TargetSizeTimeout=1000
IFS.SERVICE.CONNECTIONPOOL.READONLY.MaximumSizeTimeout=10000
These are values from "SmallServiceConfiguration" which we are using. Since we have web based application there might be time when some operations are executing concurrently. Do you think that has anything to do with Node being hung? .
We haven't tried using "medium" / "Large" service configuration yet , but are thinking to change it to "medium" . Should we do the change at the middle layer where we run our webapplication or at Node layer ?
Let me know your responses and we will try it and see whether that solves the problem.
Thanks