[omniORB] rendezvouser error
John Chesshir
john.chesshir at mainstream-tech.com
Fri Oct 28 14:28:40 BST 2005
It's been a few months since I sent this last message, but I
thought I should follow it up. Our client did eventually upgrade to
Redhat Advance Server. It really seems that the old redhat threading
model was simply not sufficient.
Much thanks to Duncan and all others who contributed to helping
track down the problem.
John Chesshir
-----Original Message-----
From: omniorb-list-bounces at omniorb-support.com
[mailto:omniorb-list-bounces at omniorb-support.com] On Behalf Of John
Chesshir
Sent: Monday, July 25, 2005 1:37 PM
To: Duncan Grisby
Cc: omniorb-list at omniorb-support.com
Subject: RE: [omniORB] rendezvouser error
OK. I have stepped into the omnithread code. (Reminder, I am
on version 4.0.5 on Red Hat Linux 7.0) The debugger went into posix.cc,
and threw an error on line 591, which calls the pthread_create function
and hands the return code to the THROW_ERRORS macro.
As I stepped through the code, I gathered this information about
what the precompiler had done with the macros:
1. PthreadDraftVersion != 4 or 8
2. __VMS is not defined
3. PthreadSupportThreadPriority is not defined
4. __linux__ is defined
Given #2 and observing execution, and looking at lines 100 to 113 of
posix.cc, I can only conclude that one of the last two of the three
THROW_ERRORS macros actually got used, both of which throw when the
return code is not equal to zero.
I was able to catch the return code of the pthread_create
function. It was 1. Looking online, I found this website
http://www.wlug.org.nz/ErrorMessages
which indicates 1 equals the error message EPERM, and according to this
website
http://www.opengroup.org/onlinepubs/007908799/xsh/pthread_create.html
EPERM coming from pthread_create means, "The caller does not have
appropriate permission to set the required scheduling parameters or
scheduling policy."
We tried running the server with root access to see if getting
past "permissions" would solve the problem. Instead, we just got a
segmentation fault.
At this point, I fully realize that this is a Red Hat 7.0 issue
and not omniORB, but I'm just wondering if anybody has faced this
problem. Red Hat has stopped supporting version 7.0.
Duncan, are you sure omniORB will work with Red Hat 7.0? Are
you possibly mixing it up with a later version? I hate testing you like
that, but I just want to be sure. I have heard that their threading was
not very good at that version. If only our client would upgrade!
Thanks for any help!
John Chesshir
_______________________________________________
omniORB-list mailing list
omniORB-list at omniorb-support.com
http://www.omniorb-support.com/mailman/listinfo/omniorb-list
More information about the omniORB-list
mailing list