On 2019-01-22 18:42, Kirk Wolf wrote:
I've noticed that one *some* z/OS systems that the first call to the
resolver by a Unix process takes an extra 1/2 second or more.
Does anyone have any clues as to why this would be?
I have a tiny test program that demonstrates. It calls
getaddrinfo(127.0.0.1) twice.
I've highlighted the lines in the resolver trace output below that wrap the
lost time (577 ms).
The time is lost regardless of whether I run this without the resolver
trace.
On another z/OS system, this same section (before res_init Started) takes
less than 2 ms.
Also, if I run the program multiple times from the same shell with
_BPX_SHAREAS=YES, each run demonstrates the problem.
So the lost time is per-process and not per-address space.
Thanks,
Kirk Wolf
Dovetailed Technologies
http://dovetail.com
Do you have a LOOKUP statement in your TCPIP.TCPIP.DATA (or equivalent),
and if so what is the lookup sequence?
--
Regards, Gord Tomlin
Action Software International
(a division of Mazda Computer Corporation)
Tel: (905) 470-7113, Fax: (905) 470-6507
Support: https://actionsoftware.com/support/
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
