[omniORB] Deadlock with BiDir mode

[Scott Robertson]

I'm pretty sure everyone of the threads were locked in the same space.
Next time it happens I'll double check. Anything I should be looking
for?


On Tue, 2006-04-18 at 15:03 +0100, Duncan Grisby wrote:
> On Tuesday 28 March, Scott Robertson wrote:
> 
> > I have a service running on omniORB 4.0.3 that's been dead locking. If I
> > attach gdb to the service and debug it, I see about 270 threads all
> > locked in the same place as in the following trace back. 
> > 
> > I'm wondering if it's about time for me to upgrade to the latest version
> > of omniORB. Anyone know whether there were any locking issues fixed
> > since 4.0.3? 
> 
> Sorry for the delay in coming to this...
> 
> There have been some locking related problems fixed since 4.0.3, but I'm
> not sure this is one of them. I'd certainly suggest you update to 4.0.7
> (or better 4.1.0 beta) to see if the problem still occurs before
> spending time tracking this down.
> 
> That said, in your stack traces, where any threads blocked anywhere
> other than in getBiDirServiceContext?
> 
> Cheers,
> 
> Duncan.
>