Conference humane::scheduler

1172.0. "Synchronization fails using clusteralias " by STKEIS::SVENSSON () Tue Oct 22 1996 10:59

    
    hello,
    
    We are suffering from the same problem as described in topic 917, using
    the same application, but in our case ONLY if we use clusteralias name.
    
    VMS 6.1  (DECstep v6 platform)
    DECnet/OSI 6.2-ECO1
    Scheduler 2.1B-5  (was patched from 2.1-2 in order to solve this
    problem, but unfortunately it didn't help)
    
    
    The QBR_CONTROL_JOB stays in DEP wait status, even if it looks
    as if the job on the remote cluster terminates with success. 
    
    All nodes in these two clusters have alias name defined and there
    is no decnet communication problems betwen them using alias.
    
    The ncl application sched_decnet is correct defined
    and the proxy accounts are setup with alias and exists in both
    netproxy.dat and net$proxy.dat.
    
    
    Identifiers
    
        Name                              = SCHED_DECNET
    
    	Client                            = <Default value>
        Addresses                         =
           {
              name = SCHED_DECNET
           }
    	
    	Outgoing Proxy                    = True
        Incoming Proxy                    = True
        Outgoing Alias                    = True
        Incoming Alias                    = True
        Node Synonym                      = True
        Image Name                        = SYS$SYSTEM:SCHED_DECNET.COM
        User Name                         = "SCHED$DECNET"
        Incoming OSI TSEL                 = <Default value>
        Network Priority                  = 0 
    
    
    
    Any hints will be appreciated.
    
    Per-Olof, CCS Stockholm
    
    Here is debug from the remote cluster
    
    
we woke up!
got mbx msg '>>1643    '
02:28 PM  processing record #  1643  status= S   request=  
 Now=22-OCT-1996 14:28:12.84   job_sched_time=22-OCT-1996 14:28:12.52
Job  1643  is scheduled for the past - check pre-requisites
All Deps must have completed with success later than 
22-OCT-1996 14:27:56.55
calling RUN_TASK to run job  1643 
Load balancing to STKBAT
cluster_broadcast:---node=STKBAT msg=%%1643
timer flag was clear
timer not expired. No earlier event to set.
sleeping
    
we woke up!
got mbx msg '%%1643    '
02:28 PM  processing record #  1643  status= S   request= N
vss$get_next_start_time: 1  cstat= 1  next=NEVER                  
Running Job  1643  PID=6020C82B Count= 3  Priority= 4 
timer flag was clear
timer not expired. No earlier event to set.
sleeping
we woke up!
job #  1643  finished.... count=  2 
exit status of job was 10000001
Job  1643  NORETAIN on success, deleting
cluster_broadcast:---node= msg=CWJ
timer flag was clear
timer not expired. No earlier event to set.
sleeping
got mbx msg 'DEBUG OFF '
Setting Debugging OFF
    
    
Debug from the "local" cluster from where the jobs are initated
    
we woke up!
job #  258  finished.... count=  0 
exit status of job was 00000001
 0  remote nodes care about job  258 
02:28 PM  processing record #  258  status= S   request=  
 Now=22-OCT-1996 14:28:22.66   job_sched_time=31-DEC-5999 00:00:00.00
job  258  is scheduled for the future
02:28 PM  updated    record #  258  status= S    request= 
Found 0  local jobs depending on :: 258 
cluster_broadcast:---node= msg=CWJ
timer flag was clear
timer not expired. No earlier event to set.
sleeping
we woke up!
got mbx msg 'DEBUG OFF '
Setting Debugging OFF
    
    
No information here from the QBR_CONTROL_JOB which had id 306.

    Hate to say it, but install scheduler V2.1B-9.