Re: [Condor-users] jobs stuck in queue

[Fabricio Cannini <fcannini@xxxxxxxxx>]

Em terça-feira 30 agosto 2011, às 16:31:41, Matthew Farrellee escreveu:
> On 08/30/2011 03:20 PM, Fabricio Cannini wrote:
> > Em quarta-feira 24 agosto 2011, às 14:48:49, David Brodbeck escreveu:
> >> On Tue, Aug 23, 2011 at 9:21 PM, Koller, Garrett

If it helps, this is the ouput of NegotiatiorLog of the most recent attempt to 
run a mpi ( mpich2 ) program :


08/30/11 19:48:01 ---------- Started Negotiation Cycle ----------
08/30/11 19:48:01 Phase 1:  Obtaining ads from collector ...
08/30/11 19:48:01   Getting all public ads ...
08/30/11 19:48:01   Sorting 32 ads ...
08/30/11 19:48:01   Getting startd private ads ...
08/30/11 19:48:01 Got ads: 32 public and 24 private
08/30/11 19:48:01 Public ads include 2 submitter, 24 startd
08/30/11 19:48:01 Phase 2:  Performing accounting ...
08/30/11 19:48:02 Phase 3:  Sorting submitter ads by priority ...
08/30/11 19:48:02 Phase 4.1:  Negotiating with schedds ...
08/30/11 19:48:02   Negotiating with DedicatedScheduler@master at 
<127.0.0.1:9680>
08/30/11 19:48:02 0 seconds so far
08/30/11 19:48:02 condor_read() failed: recv() returned -1, errno = 104 
Connection reset by peer, reading 5 bytes from schedd 
DedicatedScheduler@master.
08/30/11 19:48:02 IO: Failed to read packet header
08/30/11 19:48:02     Failed to get reply from schedd
08/30/11 19:48:02   Error: Ignoring submitter for this cycle
08/30/11 19:48:02  negotiateWithGroup resources used scheddAds length 0 
08/30/11 19:48:02 ---------- Finished Negotiation Cycle ----------