Before installing the Cluster Edition:
Install operating system patches, if required.
Create a “sybase” account on your system to perform all installation tasks. This account should own all devices and files, and must have permission to read and write all devices the cluster uses.
Create all devices with this “sybase” account.
Log in to the machine as the “sybase” user.
Maintain consistent ownership and privileges for all files and directories. A single user—the Sybase System Administrator with read, write, and execute permissions—should perform all installation, upgrade, and setup tasks.
If you are using a local account local to just one physical box then create the sybase user on each machine and make sure it has the same user ID.
Decide where the Adaptive Server software will be installed. Make sure there is sufficient available disk space. There cannot be any spaces in the path name of the directory.The $SYBASE location must be on a shared file system, accessible from all cluster nodes using the same path.
Verify that the operating system meets the version-level, RAM, and network protocol requirements for your platform.
Verify that your network software is configured.
The Cluster Edition implicitly requires that your network be configured for the nodes you plan to include in the cluster.
If you are having connection problems, or to verify your network configuration, ping the host.
Adjust the operating system shared memory parameter for your platform
To adjust shared memory segments of the operating system, add the following line to the configuration file /etc/system, where x is the number of shared memory segments:
set shmsys:shminfo_shmseg=x
The method of controlling system resources is to add an entry to the file /etc/project such as in the following example, where project.max-shm-memory=(privileged,17179869184,deny) is the name for the project parameter:
project-sybase:200:For use by Sybase:sybase:sybase: project.max-shm-memory=\(privileged,17179869184,deny)
privileged – is a threshold value on a resource control that constitutes a point at which local actions can be triggered or global actions such as logging in to a machine can occur. Three privilege levels exist:
basic – can be modified by owner of the calling process
privilege – can be modified by privileged callers. In specifying the threshold value of “privileged,” you can use the abbreviation “priv,” such as in the following, where “17179869184” is the threshold value (16GB) on the resource control:
project.max-shm-memory=(priv, 17179869184,
deny)
system – is fixed for the duration of the operating system instance
You can set setting project.max-shm-memory while the system is running by using the prctl command. The rctladm command can be used to set things permanently.
The default value for project.max-shm-memory is 25% of the physical memory on the system. The maximum value is UINT64_MAX, which works out to 18446744073709551615 bytes, so essentially this is limited only by the size of physical memory.
deny – means attempts to use more than 16GB are denied.
Adjust shared memory segments. Do this if Adaptive Server can not get large memory as a single segment or if backup server stripes fail due to insufficient segments.
Depending on the number and types of devices you use for backup (dump) and recovery (load), you may need to adjust the shared memory segment parameter in the operating system configuration file to accommodate concurrent Backup Server processes. The default number of shared memory segments available for process attachments is 6.
Adaptive Server may allocate shared memory segments after start-up if any reconfiguration through sp_configure requires additional memory. You may need to account for these additional segments. Allocate the maximum memory you will make available to Adaptive Server, by using the allocate max shared memory configuration parameter. See the System Administration Guide for more information.
The operating system shared memory default for most Linux releases is 32MB. The minimum required by Adaptive Server is 64MB for a default server with 2K pages. A higher value is required if you plan to increase the Adaptive Server total memory.
Use the sysctl(8) method to check and adjust the operating system shared memory parameter.
To check the current shared memory size, enter:
# /sbin/sysctl kernel.shmmax
To adjust the shared memory size at runtime, where nnn is 1073741824 for 1GB of shared memory:
# /sbin/sysctl -w kernel.shmmax=nnn
To guarantee that this value is applied every time the system is started, edit the /etc/sysctl.conf file.
On Linux RHEL Update4.0 and greater, Adaptive Server running on multiple engines requires the security feature Exec-Shield to be disabled.
To disable Exec-Shield:
Add the following lines in /etc/sysctl.conf:
kernel.exec-shield=0 kernel.exec-shield-randomize=0
Enter the following for the action to take effect:
/sbin/sysctl -P
For additional information, see the Red Hat Web site.
If the server fails to restart after you have adjusted the SHMMAX parameter, you may also need to increase the value of another Kernel parameter, SHMALL, which is the maximum amount of shared memory that can be allocated. You can modify its value in the /etc/sysctl.conf file.
The syntax for increasing SHMALL is:
# /sbin/sysctl -w kernel.shmall=nnn
