Hang In Utl_tcp While Opening Connection To Unreachable Host (Doc ID 456954.1)

Last updated on SEPTEMBER 26, 2011

Applies to:

Oracle Net Services - Version: 10.2.0.2.0 and later   [Release: 10.2 and later ]
PL/SQL - Version: 10.2.0.2 and later    [Release: 10.2 and later]
Information in this document applies to any platform.

Symptoms

UTL_TCP.OPEN_CONNECTION procedure opening TCP connections to a series of hosts in the same network hangs intermittently without never returning when for any reason the remote host is not reachable.

Issue only reproduces in RAC environments.

Changes

truss output from hanging session:


connect(23, 0x1065E6280, 16, 1) (sleeping...)
455.0277 Received signal #14, SIGALRM, in connect() [caught]
...
connect(23, 0x1065E6280, 16, 1) (sleeping...)
460.0377 Received signal #14, SIGALRM, in connect() [caught]

On RAC, there is a special routine registering a timer and the default timeout for timer is defined as 5 seconds. From truss output, it is realized that this internal timer indefinitely restarts the connect() function every 5 seconds.
Setting _skgxp_udp_timed_wait_seconds to 10 changes the interval between the SIGALRM signals in the truss output.

Cause

Sign In with your My Oracle Support account

Don't have a My Oracle Support account? Click to get started

My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms