[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] FW: Linux computers cannot join WindowsCentralManager



GT wrote:
>> I had similar probolem before, you probably can try use 
>> UPDATE_COLLECTOR_WITH_TCP = True on central manager.

I think this is the solution to a different problem. If
your Windows boxes work OK for a while and then gradually
"drop out" of the pool, then it can be due to lost UDP 
packets in certain network situations. The above is one
solution.

Alex wrote
> >We have set up a windows XP computer as a central manager with its 
> >respective pool, we also setup another windows system as a Condor 
> >ViewServer , joining windows systems to the pool doesn't 
> seem to be a 
> >problem but when we want to join Linux systems we cannot do 
> it. We had 
> >reviewed all settings on the Linux systems and all seem to 
> be alright, 
> >we have no errors to reports, we had reviewed all logs and we cannot 
> >find any errors that can tell us what to troubleshoot. Windows and
> Linux
> >systems can ping each others. 

This sounds like it may be a firewall problem. Being able to "ping"
whilst a good
start isn't sufficient. Do you have the appropriate udp and tcp ports
open in both
directions between
* first of all: the central node (neg+coll) and all other nodes?
* secondly: all potential submit with all potential execute nodes (and
vice versa)

If not that, it could be something to do with the names. Maybe turn on
some debugging
(in the Collector for instance) and try connect a single linux and see
if anything gets through.

Cheers

JK