Starting the cluster

Start the cluster by starting all the instances associated with the cluster on each node in the cluster.

You can start a cluster:

Normally, after a graceful shutdown. See “Starting the cluster after a normal shutdown”.
After a system failure. See “Starting the cluster after a system failure”.

Sybase recommends that you do not alter the runserver files used for normal cluster startup.

Starting the cluster after a normal shutdown

Start each cluster instance from the node on which it runs:

For example, start the “ase1” instance:
```
startserver -f $SYBASE/$SYBASE_ASE/install/RUN_ase1
```
Log in to each node on which another instance will run, and execute startserver.

For example, to start the “ase2” instance on “blade2”:
1. On “blade2,” move to the $SYBASE directory.
2. Issue the startserver command:
```
startserver -f $SYBASE/$SYBASE_ASE/install/RUN_ase2
```

Starting the cluster after a system failure

If the cluster goes down because of a system failure and you have enabled the automatic cluster takeover configuration parameter follow step 1. If you have not enabled the automatic cluster takeover configuration parameter follow step 2.

Set it to the default, which is 1. The instance attempting to connect to the cluster restarts the cluster, acting as the cluster coordinator and recovering any databases. See “Configuring the cluster for automatic cluster takeover” for information about configuring the cluster to automatically restart.
If you do not enabled automatic cluster takeover, you must reboot the cluster with the dataserver . . . --cluster_takeover parameter. For example, this example adds the -cluster_takeover parameter to the runserver file that originally started the cluster:
```
$SYBASE/ASE-15_0/bin/dataserver \
--quorum_dev=/dev/raw/raw11\
--instance=ase1\
--cluster_takeover
```
Once the cluster is running, restart all the instances as you regularly would.