Tuxedo - on Solaris 8: signal delivery issue may affect networked components

(Doc ID 774110.1)

Last updated on NOVEMBER 04, 2016

Applies to:

Oracle Tuxedo
Information in this document applies to any platform.
Information in this document applies to any platform
***Checked for relevance on 24-01-2011***

Goal

On Solaris 2.8 (as well as 2.7), a bug in the Kernel may cause delays in some processes because a network signal may
not be delivered.

The networked components in Tuxedo (/WS, /Domains, BRIDGE, Jolt) are designed in a highly asynchronous mode to enhance
performance, and rely on the fact that the OS delivers a SIGIO or SIGPOLL signal to the corresponding process (WSH,
GWTDOMAIN, BRIDGE or JSH) whenever a network event happens: typically a request message coming in from a client
process.

On Solaris 2.8, however, a kernel bug will sometimes cause those networked components not to receive the network
signal, possibly leading to a delay. This is mostly apparent with the WSH or JSH processes, that would simply sleep on
a msgrcv() system call for as long as SCANUNIT * SANITYSCAN all the while messages are coming in from the network side
(as seen with truss and snoop traces).

Solution

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