0: *Recently I was charged with creating a two node Veritas Cluster
1: *server. Everything in VCS resolves around this config file.
2: *Since all of the binaries in VCS are written in Perl and with
3: *the increasing popularity of this product I thought some monks
4: *might find this useful. This configuration fails over four
5: *resources, Apache, Sendmail, SMMS and the RAID. When any
6: *of the apps fail Veritas umounts /var/md and mounts it on
7: *the other system and then starts up the apps. you must write
8: *a small monitoring script for each app and place it in
9: */opt/VRTSvcs/bin. The monitoring script simply needs to
10: *return "online" or "offline" (pgrep works well)
11: *Anyways I hope someone finds this useful
12: *the config file is pretty easy to read. Cheers!!!
13:
14:
15:
16: include "types.cf"
17:
18: cluster Mail (
19: UserNames = { admin = "cDRpdxPmHpzS." }
20: CounterInterval = 5
21: Factor = { runque = 5, memory = 1, disk = 10, cpu = 25,
22: network = 5 }
23: MaxFactor = { runque = 100, memory = 10, disk = 100, cpu = 100,
24: network = 100 }
25: )
26:
27: system wgmail1a
28:
29: system wgmail1b
30:
31: snmp vcs (
32: TrapList = { 1 = "A new system has joined the VCS Cluster",
33: 2 = "An existing system has changed its state",
34: 3 = "A service group has changed its state",
35: 4 = "One or more heartbeat links has gone down",
36: 5 = "An HA service has done a manual restart",
37: 6 = "An HA service has been manually idled",
38: 7 = "An HA service has been successfully started" }
39: )
40:
41: group mailgroup (
42: SystemList = { wgmail1a, wgmail1b }
43: AutoStartList = { wgmail1a }
44: )
45:
46: Application Apache (
47: User = root
48: StartProgram = "/etc/init.d/apache start"
49: StopProgram = "/etc/init.d/apache stop"
50: MonitorProgram = "/opt/VRTSvcs/bin/monitor_scripts/monitor_http"
51: )
52:
53: Application SMMS (
54: User = root
55: StartProgram = "/etc/init.d/smmsd start"
56: StopProgram = "/etc/init.d/smmsd stop"
57: MonitorProgram = "/opt/VRTSvcs/bin/monitor_scripts/monitor_smms"
58: )
59:
60: Application sendmail (
61: User = root
62: StartProgram = "/etc/init.d/sendmail start"
63: StopProgram = "/etc/init.d/sendmail stop"
64: MonitorProgram = "/opt/VRTSvcs/bin/monitor_scripts/monitor_sendmail"
65: )
66:
67: IP ip_eri0 (
68: Device = eri0
69: Address = "199.6.154.34"
70: NetMask = "255.255.255.0"
71: )
72:
73: Mount the_raid (
74: MountPoint = "/var/md"
75: BlockDevice = "/dev/dsk/c2t5d1s0"
76: FSType = ufs
77: MountOpt = logging
78: )
79:
80: NIC nic_eri (
81: Device = eri0
82: NetworkType = ether
83: )
84:
85: sendmail requires the_raid
86:
87:
88: // resource dependency tree
89: //
90: // group mailgroup
91: // {
92: // Application Apache
93: // Application SMMS
94: // IP ip_eri0
95: // NIC nic_eri
96: // Application sendmail
97: // {
98: // Mount the_raid
99: // }
100: // }
In reply to Veritas Cluster Server Failover by JSchmitz
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |