This article will help you with a proper restart of the CML server that is part of CML cluster. In addition the article describes the steps you shall take if you want to take one CML in a cluster offline for a longer period of time. Procedure mainly describes the steps in a CML cluster environment but it can be used also for environment with just one CML node.

Article is divided into two major parts:

1) How to stop the YSoft SafeQ server

use this procedure in case that you need to plan the server downtime e.g. for the server maintenance.

2) How to start the YSoft SafeQ server

use this procedure in case that you need to start the server after the planned downtime. This procedure can be used also in case that server has unexpectedly gone offline and you need to start it up again (e.g. after one day).

1) How to stop the YSoft SafeQ server

In these instructions, the server you want to restart/temporarily stop is the A server; the other servers are B and C servers.

WARNING:

If the server A is the master node (1st installed node), some additional functions such as LDAP replication, displaying of the web reports or sending of statistical data to the CRS server may not be available during the downtime.
If the server A is the master node (1st installed node), keep the downtime as short as possible (maximally 7 days). Otherwise some statistical data may not be generated.
Configuration in the system settings cannot be changed during the downtime period of any CML node in the cluster.

1	Make sure that other nodes (B and C) are working correctly and that their job reception or reception of connections from the terminals is not STOPPED. You can verify this in the administration interface (YSoft SafeQ Web Interface => System => System Information => Services). All the below mentioned services must have the status RUNNING: TCP/IP Print Server on CML (SafeQ Port) TCP/IP Print Server on CML (LPD) Server Terminal UDP Identificator SafeQ Terminal Listener Enterprise Server Balance Communicator Enterprise Server Synchronization
2	Verify that there are no active locks on the A node (YSoft SafeQ Web Interface => System => System Information => Services). If the locks are held wait until all operations are finished, users have logged off the terminals and locks are removed. WARNING: The system can be stopped without this step. However a sudden termination of the services may damage the jobs that are currently processed by the active sessions.
3	Stop reception of new jobs and connections from terminals on the node A you plan to stop: YSoft SafeQ Web Interface => System => System Information => Services => press Stop service at following objects: TCP/IP Print Server on CML (SafeQ Port) TCP/IP Print Server on CML (LPD) SafeQ Terminal Listener
4	Move the print jobs from the CML that shall be stopped to another node in the cluster that will remain active: Perform this step only if: cluster-readSharedFolderJob in the system settings is disabled you plan to keep the CML server offline for the longer period of time (e.g. 10 hours) and you wish to have the already spooled jobs available you are stopping the CML server that is part of the CML cluster If you omit this step, the system will continue to work but some print jobs sent to the node B will not be available for print. Users trying to release such jobs may experience spooler error message. 4a. Copy the content of the spooler from the node to be restarted (A) to other node (B or C) Spooler is located by default in <safeq_home>\server\spool. 4b. Change the location of the spooled jobs in the database Open pgAdminIII or MS SQL management studio on any running CML. Connect to the SQDB5 database using the same user/password that is used by YSoft SafeQ. Do not use the account Sync. Run the following SQL query on the SQDB5 database: update smartq_jobs SET server_id=<OTHER_NODE_ID> where server_id=<STOPPED_NODE_ID> *Example:* update smartq_jobs SET server_id=3 where server_id=1 In our example server A had ID number 1 and the jobs were copied to the server C with ID number 3. The ID of servers can be found in the same database in table smartq_servers column id. All changes in the database will be automatically replicated to all the other nodes.
5	Reconfigure the print drivers and terminals to connect to the node that will remain active (not the one you are stopping) Perform this step only if: you have a CML cluster you plan to keep the CML server offline for the longer period of time (e.g. 10 hours) and you wish to have the already spooled jobs available 5a. Reconfigure all print drivers (or YSoft SafeQ clients) to connect to the active server You can use the procedure from guide CML - Printer settings in case of one CML node failure to sustain the printing functionality. Otherwise users might experience issues with the job printing from the workstation to the YSoft SafeQ server. 5b. Reconfigure HW terminals: If you are using hardware terminals, make sure that they are pointing to the new IP address. Otherwise users will be unable to authenticate. 5c. Reinstall all embedded Terminals: Reinstall all embedded terminals that shall be connecting to the server with the new IP address. Otherwise users will be unable to authenticate. This step can be skipped if you are using Embedded terminal failover.
6	Open the Services window (e.g. Start => Run => services.msc) and stop all YSoft SafeQ services YSoft SafeQ CML YSoft SafeQ CML DBS YSoft SafeQ LDAP Replicator Service YSoft SadeQ Terminal Server YSoft SafeQ Web Interface

2) How to start the YSoft SafeQ server

If one of the cluster nodes has been down for some time (maintenance, updates, error-related downtime), all the missing data will be automatically synchronized between servers once the node is connected again.
With the default settings, the synchronization speed is approx. 20,000 entries per minute, i.e. information about 4,000 jobs is replicated every minute. Therefore we recommend to connect stopped server as follows:

1	Connect server to the network outside the peak hours with YSoft SafeQ services stopped. In case that services cannot be stopped right after the connection, stop them as soon as possible except for YSoft SafeQ CML DBS service.
2	Start synchronization process by launching YSoft SafeQ CML DBS service (if it is not running)
3	Connect to the SafeQ web interface of all servers in the cluster, go to => System => System information => Services. Verify that all numbers in the Database pool priority are close to zero (synchronization is almost finished).
4	Start all the remaining SafeQ services that are stopped. In case that printers were reconfigured to connect to another node during the downtime period, make sure that printers are reconfigured to the previous settings.