Both sybcluster and the Adaptive Server plug-in automatically detect if the SF for Sybase CE is installed on the host node, and guide you to select the appropriate storage and network interconnects. When you create a cluster:
All storage must be on SF for Sybase CE managed devices, using CVM controlled volumes or SFCFS file systems, with these restrictions:
If you access CVM controlled volumes directly, you must access them as raw (character) devices, using /dev/vx/rdsk (instead of /dev/vx/dsk/)
You can place private devices added specifically for local temporary databases on local, non-SF for Sybase CE storage devices.
Do not use storage provided by non-Sybase CE versions of Veritas Storage Foundation. Do not use storage provided by SF for Sybase CE unless you configure the Cluster Edition for vcs membership mode. Adaptive Server cannot perform VxFEN-controlled I/O fencing in native mode as database corruption may result.
Configure the VCMP socket path. By default, the VCMP uses /tmp/vcmp_socket to communicate with Veritas.
Add this to the runserver file located in $SYBASEASE-15_0/install/RUN_instance name:
export VCMP_SOCKET=new_vcmp_socket_path
For example:
export VCMP_SOCKET=/tmp/my_socket
Modify the vxfend resource:
hares -modify vxfend Arguments "%-m sybase -k new_vcmp_socket_path"
For example:
hares -modify vxfend Arguments "%-m sybase -k /tmp/my_socket"
i.e.
The Cluster Edition primary and secondary cluster interconnects must run over Veritas LLT managed networks. Determine which networks LLT uses from the /etc/llttab file. In this example, LLT runs over the eth0 and eth1 networks:
[admin@sdcc2 ~]$ cat /etc/llttab set-node sdcc2 set-cluster 53 link eth0 eth-00:19:b9:b0:73:13 - ether - - link eth1 eth-00:19:b9:b0:73:15 - ether - -
Cluster Edition interconnects require that you configure
IP addresses for the links. Because LLT does not require IP addresses,
you may need to configure the interconnects separately. See your
operating system documentation.
Determine the IP address assigned to a given network using the ifconfig command. You can use this address to configure either the primary or secondary cluster interconnect. In this example, the interface eth0 has an IPv4 address of 10.22.104.141:
[admin@sdcc2 ~]$ /sbin/ifconfig -a eth0 eth0 Link encap:Ethernet HWaddr 00:19:B9:B0:73:13 inet addr:10.22.104.141 Bcast:10.22.104.255 Mask:255.255.255.0 inet6 addr: fd77:55d:59d9:168:219:b9ff:feb0:7313/64 Scope:Global inet6 addr: fe80::219:b9ff:feb0:7313/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:19710010 errors:0 dropped:0 overruns:0 frame:0 TX packets:11707065 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:7506123489 (6.9 GiB) TX bytes:1839949672 (1.7 GiB) Interrupt:169 Memory:f8000000-f8012100
Reset the cluster to vcs mode after you create the cluster.
