Should we disable Solaris using Unix Domain Sockets in the regression test?
Hi all,
Did anyone ever figure out why Solaris boxes had those random failures
during the regression tests?
If no, then should we ensure the regression test on Solaris run's over
TCP instead (like BeOS and QNX do)? To do this is a one-line fix in the
shell script.
Regards and best wishes,
Justin Clift
--
"My grandfather once told me that there are two kinds of people: those
who work and those who take the credit. He told me to try to be in the
first group; there was less competition there."
- Indira Gandhi
Justin Clift <justin@postgresql.org> writes:
Did anyone ever figure out why Solaris boxes had those random failures
during the regression tests?
It should be better now that listen() is being called with a larger
number. The basic problem is that Solaris actually honors the
listen() backlog argument, and if more than that many Unix socket
clients attempt to connect simultaneously, some are rejected
immediately.
Ian
Cool.
:)
Regards and best wishes,
Justin Clift
Ian Lance Taylor wrote:
Justin Clift <justin@postgresql.org> writes:
Did anyone ever figure out why Solaris boxes had those random failures
during the regression tests?It should be better now that listen() is being called with a larger
number. The basic problem is that Solaris actually honors the
listen() backlog argument, and if more than that many Unix socket
clients attempt to connect simultaneously, some are rejected
immediately.Ian
--
"My grandfather once told me that there are two kinds of people: those
who work and those who take the credit. He told me to try to be in the
first group; there was less competition there."
- Indira Gandhi