High To 100% CPU Utilization By oidldapd After Applying OID 10.1.4.3 Patchset If EUS Or Other SASL Binds Used (Doc ID 825165.1)

Last updated on SEPTEMBER 17, 2020

Applies to:

Symptoms

Oracle Internet Directory (OID) 10.1.4.0.1 was operating normally without problems.
OID has also been properly tuned as per <Note 458703.1> already.

Since applying Patchset 10.1.4.3, oidldapd processes consume high to 100% cpu.

The oidldapd instances that have the problem service mostly Enterprise User Security (EUS) or other SASL connections.

The OID log for the instance number in question shows only the following error:

Note: The OS error number varies by platform, i.e.:

For OID Server on HP:
2007/06/05:10:41:12 * ServerDispatcher :2 * sgslufread: Hard error on read, OS error = 232

For OID Server on Solaris:
2007/06/13:08:36:26 * ServerDispatcher :2 * sgslufread: Hard error on read, OS error = 131

For OID Server on Linux:
2008/07/01:02:41:17 * ServerDispatcher :1 * sgslufread: Hard error on read, OS error = 104

For OID Server onWindows:
2008/07/01:02:41:17 * ServerDispatcher :1 * sgslufread: Hard error on read, OS error = 10054

These OS errors (232, 131, 104, 10054) all mean: Connection reset by peer

Changes

Applied Patchset 10.1.4.3.0.

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.