I am trying to execute a job from the Console, but get the following exception when the job is executed-
“Exception thrown when trying to communicate with Job Server:XXXXXXv007:3500. Socket is closed”. I do not have an option to execute the job from Designer due to firewall issues. I was able to telnet to the job server port using command prompt from the Client.
Please let me know if this is related to network configuration between Console and Job Server? and any steps to resolve this?
Is that the correct port number? Sometimes the port number defined for the job server (using Server Manager) and the port number recorded in the repository table (al_machine_info) are not in agreement.
you can’t see the status of DS Job Server in CMC, for DS Job Server check if you have started the Data Services Service (Windows service) also check for the existence of al_jobserver process in the task manager
check the DS Job server log fine in %LINK_DIR%\log<JobServerName>\server_eventlog_.txt file for any errors
For the repository in which the ETL job resides please post the output of table AL_MACHINE_INFO. You may mask the actual machine name when you post the results to the forum.
The values in that table must match the current configuration of the job server (machine name/ip address, port numbers, etc).
eganjp,
Here is the output from AL_MACHINE_INFO
SEQNUM || SERVER_NAME || MACHINE_NAME || PORT
44 || Administrator || XXXl08v007 || 8080
36 || Administrator || XXXl08v007.XXX.com || 8080
1 || RepoManager || ||
2 || JobServer_1 || XXXl08v007 || 3500
Couple of my observations-
XXXl08v007, XXXl08v007.XXX.com are the different host names to same IP address. To log into CMS and DS Console, we use XXXl08v007XX7 that is a third host name to the same IP address
Why is the RepoManager values blank for machine name and port?
XXXl08v007 is the name that the Management Console web server/app server is going to use to try to talk to the job server. The port number is also used. If either of these are wrong then the Management Console can’t start the job and you get an error. If the Management Console and the job server are on the same machine you could change the MACHINE_NAME value to LocalHost I belive. I would only do this to troubleshoot. I don’t normally leave it that way.
If you have other repositories on the same job server that are working then compare the values in AL_MACHINE_INFO and make adjustments as needed.
It’s possible that the information in AL_MACHINE_INFO is completely correct. But it’s where I would start.