[torqueusers] host not authorized

Brock Palen brockp at umich.edu
Fri Apr 25 12:19:59 MDT 2008


wait, you may be right,

qmgr -c 'l s pbs_version'
         pbs_version = 2.4.0

Version 2.4  would that be the snapshot?
Is gssapi rolled into the regular torque snapshots now?

Brock Palen
www.umich.edu/~brockp
Center for Advanced Computing
brockp at umich.edu
(734)936-1985



On Apr 25, 2008, at 2:16 PM, Brock Palen wrote:
> I need to double check with the guy who started this project, but  
> we are using the GSSAPI branch,  and we might be using a snapshot  
> from 2.2 well before this bug.
>
> Brock Palen
> www.umich.edu/~brockp
> Center for Advanced Computing
> brockp at umich.edu
> (734)936-1985
>
>
>
> On Apr 25, 2008, at 1:44 PM, Steve Snelgrove wrote:
>> I am very sorry to have introduced this problem.  It had to do  
>> with an attempt at dealing with a pbs_server with multiple  
>> Ethernet interfaces where they wanted the high level server  
>> traffic to go over one interface and the mom traffic to go over  
>> the other.  In attempting to solve this (done back when 2.3 was  
>> the developement branch) the code to insert the pbs_server into  
>> the okclients list was removed.  I have replaced this code  
>> yesterday and fixed the problem in 2.3 and 2.4(trunk).
>>
>> http://www.clusterresources.com/downloads/torque/snapshots/ 
>> torque-2.3.1-snap.200804241117.tar.gz
>> http://www.clusterresources.com/downloads/torque/snapshots/ 
>> torque-2.4.0-snap.200804241119.tar.gz
>>
>> Again, I am very sorry to have caused this problem and am thankful  
>> for the many good eyes who have correctly pointed out the place in  
>> the code where the error was observed.  Thanks.
>>
>>
>> Brock Palen wrote:
>>> I am having trouble setting up torque on a host.  I have done  
>>> this many times, but I cant figure out why the mom will not  
>>> accept commands from the server.
>>>
>>> When I try to do 'qrun'  on the server,  (I am a manager and  
>>> operator)
>>>
>>> qrun: Execution server rejected request  
>>> REJHOST=nyx190.engin.umich.edu MSG=cannot send job to  
>>> nyx190.engin.umich.edu, state=PRERUN 1.gridlock.engin.umich.edu
>>>
>>> I also can't use momctl from the server:
>>>
>>> gridlock% /usr/local/sbin/momctl -h nyx190.engin.umich.edu -d 3
>>> ERROR:    query[0] 'diag3' failed on nyx190.engin.umich.edu  
>>> (errno: 0:5)
>>>
>>>
>>> On the mom,  I see the following:
>>>
>>> 04/25/2008 13:11:43;0008;   pbs_mom;Job;process_request;request  
>>> type GSSForwardCreds from host gridlock.engin.umich.edu rejected  
>>> (host not authorized)
>>>
>>> Now this is the GSSAPI branch, but I don't think GSSAPI has  
>>> anything to do with this.  Looking in server/proccess_request.c,   
>>> the code looks like this:
>>>
>>>   if (!tfind(svr_conn[sfds].cn_addr,&okclients))
>>>    {
>>>     sprintf(log_buffer,"request type %s from host %s rejected  
>>> (host not authorized)",
>>>       reqtype_to_txt(request->rq_type),
>>>       request->rq_host);
>>>
>>>     log_record(
>>>      PBSEVENT_JOB,
>>>       PBS_EVENTCLASS_JOB,
>>>       id,      log_buffer);
>>>
>>>     req_reject(PBSE_BADHOST,0,request,NULL,"request not  
>>> authorized");
>>>
>>>     close_client(sfds);
>>>     return;
>>>     }
>>>
>>>
>>> We used to have to set /etc/hosts.equiv etc,  to make things  
>>> work, but have not had to do that for some time.  I cant even use  
>>> momctl,
>>>
>>> Thanks guys.
>>>
>>> Brock Palen
>>> www.umich.edu/~brockp
>>> Center for Advanced Computing
>>> brockp at umich.edu
>>> (734)936-1985
>>>
>>>
>>>
>>> _______________________________________________
>>> torqueusers mailing list
>>> torqueusers at supercluster.org
>>> http://www.supercluster.org/mailman/listinfo/torqueusers
>>>
>>
>>
>>
>
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers
>
>



More information about the torqueusers mailing list