Not sure. The message was:
OOB: Connection to HNP lost
I have a bigger problem now though. As of rc1, terminating a job no
longer works. I'll try rc2 and let you know if the problem persists.
Since the API for terminate changed recently, I updated the code to
replicate what happens in orte
Hopefully. Was it IOF-related?
The error was that some in-flight IOF fragments (meaning that they
had been read from the local source and were in the process of being
sent across OOB) could be incorrectly removed from the list, later
causing either a segv in production builds or, more reli
Thanks. I was seeing an error when I shut down orted. Sounds like
it's now fixed...
Greg
On Mar 9, 2007, at 5:25 PM, Jeff Squyres wrote:
- An IOF race condition in the shutdown of the orted
- Some sm btl fixes
- Patch to change Libtool 2.0 libltdl's behavior with regards to
lt_dlopen'ing DS
- An IOF race condition in the shutdown of the orted
- Some sm btl fixes
- Patch to change Libtool 2.0 libltdl's behavior with regards to
lt_dlopen'ing DSOs
On Mar 9, 2007, at 7:54 PM, Greg Watson wrote:
What changed between rc1 and rc2?
Greg
On Mar 9, 2007, at 1:50 PM, Tim Mattox wrote:
What changed between rc1 and rc2?
Greg
On Mar 9, 2007, at 1:50 PM, Tim Mattox wrote:
Hi All,
The second release condidate of v1.2 is now up on the website:
http://www.open-mpi.org/software/ompi/v1.2/
Please run it through it's paces as best you can.
--
Tim Mattox - http://homepage.mac.com/tma
