Responding to unattended ORS server reboot

• Near Roaming Group with 2 members

  • follow Whole NRG cluster restart section.

• Near Roaming Group with 3 or more members

  • Follow Restarting single ORS section

  • If there is a need to reboot more than one server at the same time, follow Whole NRG cluster section.

Restarting single ORS

• Near Roaming Group with 2 members

  • follow Whole NRG cluster restart section.

• Near Roaming Group with 3 or more members

  • Only one server can be going through the procedure below. Once completed, another server can be restarted. It is forbidden to restart 2 or more ORS servers at the same time. Should that be required, proceed with Whole NRG cluster restart.

  • Restart single ORS only when NRG is properly established. Before restart, verify that NRG is properly established on all ORS servers according to Verification that NRG is properly established section.

  1. Stop all three ORS services in no particular order.

  2. Delete cache (unless there is a specific guideline prepared by YSoft).

     1. If cache deletion is undesirable, then follow "ORS service only restart" guideline.

     2. If orsCacheRecovery property is disabled, manually delete all jobs from spooler while deleting the cache
        

  3. Start service YSoft SafeQ ORS.

     1. Do not start YSoft SafeQ ORS Web Service manually, proceed to next step. YSoft SafeQ ORS Web Service is started automatically.

  4. Open ors.log located in SafeQORS\logs and wait until all data (users, jobs, devices, queues, terminals,... ) are downloaded.

     1. In case orsFailoverLockManager property is enabled and appropriate license with " ORS failover and load balancing" feature is used , search for Download entities finished from other ORSes in NRG. Note: This can take up to 15 minutes, depending on number of users and jobs.

     2. In case license without " ORS failover and load balancing" is used, search for End of processing of GetNewJobsByUsersResponseMessage. Note: Message appears in ors.log approximately after time defined in refreshRoamingJobCronRule property.

  5. Start service YSoft SafeQ Terminal Server service. ORS should be excluded from authentication until this step.

  6. At this point the ORS is fully operational.

  7. Proceed as described in section Verification that NRG is properly established below.
     

Whole NRG cluster restart

Whole NRG cluster restart means that all ORS servers in that NRG are restarted. The procedure is as follows:

You can proceed with the following steps on each ORS at the same time. ORSes can be starting up in parallel, but it is good practice to start ORS servers one by one and start next ORS server when previous ORS server is fully operational (as described in step 6). Starting ORS servers one by one decreases load on CML.

1. On each server in the NRG stop all three ORS services in no particular order

2. Delete cache on all ORS servers

   • If orsCacheRecovery property is disabled, manually delete all jobs from all spoolers.

3. Start service YSoft SafeQ ORS.

   1. Do not start YSoft SafeQ ORS Web Service manually, proceed to next step. YSoft SafeQ ORS Web service is started automatically.

4. Open ors.log located in SafeQORS\logs and wait until all data (users, jobs, devices, queues, terminals,... ) are downloaded from distributed layer.

   • In case orsFailoverLockManager property is enabled and appropriate license with " ORS failover and load balancing" feature is used , search for Download entities finished from other ORSes in NRG. Note: This can take up to 15 minutes, depending on number of users and jobs.

   • In case license without " ORS failover and load balancing" is used, search for End of processing of GetNewJobsByUsersResponseMessage. Note: Message appears in ors.log approximately after time defined in refreshRoamingJobCronRule property.

5. Start service YSoft SafeQ Terminal Server. ORS should be excluded from authentication until this step.

6. At this point the ORS is fully operational.

Once all ORS servers are started, proceed as described in section Verification that NRG is properly established below.

Special cases

Executed upon request from Y Soft CSS.

Verification that NRG is properly established

Repeat the procedure for each ORS server in the NRG:

1. Connect to any ORS server in NRG

2. Go to directory c:\SafeQORS\bin and start file sqjconsole.bat

3. To tab Remote Process: enter <ORS_IP>:9999
   
   

4. When jConsole connects to ORS Web service go to tab MBeans

5. In menu navigate to distCache / CacheManager / "DefaultCacheManager" / CacheManager / Attributes

6. Verify that value of attribute clusterSize corresponds to number of members in NRG

   

7. Go to directory c:\SafeQORS\logs and open infinispan-server-rest.log.0 file and find last "Received new cluster view:" record. It must contain all NRG members in square brackets.
   e.g. "INFO: ISPN000094: Received new cluster view: [QA11-S085-13009|1] [QA11-S085-13009, QA11-S086-44881]" This NRG contains 2 ORS servers.

• If there is a need to reboot more than one server at the same time, follow Whole NRG cluster section.