Re: [HTCondor-users] Problem with REPLICATION_USE_SHARED_PORT

[Andrea Sartirana <sartiran@xxxxxxxxxxxx>]

I've forgot the condor version

condor-8.6.0-1.el6.x86_64

Cheers,
Andrea


On 10/03/2017 17:07, Andrea Sartirana wrote:
> Hi,
>
> at GRIF we are currently testing HAD and Replication.
> Things work just fine when declaring special ports for HAD and 
> REPLICATION. But, when setting REPLICATION_USE_SHARED_PORT to TRUE, 
> the replication service refuses to start and I see errors like these 
> in the Master log of the 2 master servers
>
> 03/10/17 17:01:26 ERROR: SharedPortEndpoint: failed to bind to 
> 15f287e5db818c2dbce9638b70a6dc044992f0be80d2dc43848c983c1fc43fa5/MASTER: 
> Address already in use
> 03/10/17 17:01:26 ERROR: Create_Process failed trying to start 
> /usr/sbin/condor_replication
> 03/10/17 17:01:26 restarting /usr/sbin/condor_replication in 265 seconds
>
> Below [1] my HAD/REPLICATION configuration.
> .... What am I doing wrong?
>
> Thanks,
> Andrea
>
> [1]
> HAD_USE_SHARED_PORT = TRUE
> REPLICATION_USE_SHARED_PORT = TRUE
> REPLICATION_LIST = lpnhe-gs9088.in2p3.fr:$(SHARED_PORT_PORT) 
> llrmpicream.in2p3.fr:$(SHARED_PORT_PORT)
> HAD_LIST = lpnhe-gs9088.in2p3.fr:$(SHARED_PORT_PORT) 
> llrmpicream.in2p3.fr:$(SHARED_PORT_PORT)
>
> HAD_CONTROLLEE          = NEGOTIATOR
> HAD_CONNECTION_TIMEOUT = 10
> HAD_USE_PRIMARY = true
>
> DAEMON_LIST = $(DAEMON_LIST) HAD REPLICATION
>
> HAD_USE_REPLICATION    = true
>
> STATE_FILE = $(SPOOL)/Accountantnew.log
>
> REPLICATION_INTERVAL                 = 300
>
> MAX_TRANSFER_LIFETIME                = 300
>
> HAD_UPDATE_INTERVAL = 300
>
> MASTER_NEGOTIATOR_CONTROLLER    = HAD
> MASTER_HAD_BACKOFF_CONSTANT     = 360