1OCF_HEARTBEAT_SG_PER(7) OCF resource agents OCF_HEARTBEAT_SG_PER(7)
2
3
4
6 ocf_heartbeat_sg_persist - Manages SCSI PERSISTENT RESERVATIONS
7
9 sg_persist [start | stop | monitor | promote | demote | meta-data |
10 validate-all]
11
13 This resource agent manages SCSI PERSISTENT RESERVATIONS. "sg_persist"
14 from sg3_utils is used, please see its documentation. Should be used as
15 multistate (Master/Slave) resource Slave registers its node id
16 ("crm_node -i") as reservation key ( --param-rk ) on each device in the
17 "devs" list. Master reserves all devices from "devs" list with
18 reservation "--prout-type" value from "reservation_type" parameter.
19
21 binary
22 The name of the binary that manages the resource.
23
24 (optional, string, default "sg_persist")
25
26 devs
27 Device list. Multiple devices can be listed with blank space as
28 separator. Shell wildcards are allowed.
29
30 (required, string, no default)
31
32 required_devs_nof
33 Minimum number of "working" devices from device list 1) existing 2)
34 "sg_persist --read-keys $device" works (Return code 0) resource
35 actions "start","monitor","promote" and "validate-all" return
36 "$OCF_ERR_INSTALLED" if the actual number of "working" devices is
37 less then "required_devs_nof". resource actions "stop" and "demote"
38 tries to remove reservations and registration keys from all working
39 devices, but always return "$OCF_SUCCESS"
40
41 (optional, string, default "1")
42
43 reservation_type
44 reservation type
45
46 (optional, string, default "1")
47
48 master_score_base
49 master_score_base value "master_score_base" value is used in
50 "master_score" calculation: master_score = $master_score_base +
51 $master_score_dev_factor * $working_devs if set to bigger value in
52 sg_persist resource configuration on some node, this node will be
53 "preferred" for master role.
54
55 (optional, string, default "0")
56
57 master_score_dev_factor
58 Working device factor in master_score calculation each "working"
59 device provides additional value to "master_score", so the node
60 that sees more devices will be preferred for the "Master"-role
61 Setting it to 0 will disable this behavior.
62
63 (optional, string, default "100")
64
65 master_score_delay
66 master/slave decreases/increases its master_score after delay of
67 $master_score_delay seconds so if some device gets inaccessible,
68 the slave decreases its master_score first and the resource will no
69 be watched and after this device reappears again the master
70 increases its master_score first this can work only if the
71 master_score_delay is bigger then monitor interval on both master
72 and slave Setting it to 0 will disable this behavior.
73
74 (optional, string, default "30")
75
77 This resource agent supports the following actions (operations):
78
79 start
80 Starts the resource. Suggested minimum timeout: 30s.
81
82 promote
83 Promotes the resource to the Master role. Suggested minimum
84 timeout: 30s.
85
86 demote
87 Demotes the resource to the Slave role. Suggested minimum timeout:
88 30s.
89
90 notify
91 Suggested minimum timeout: 30s.
92
93 stop
94 Stops the resource. Suggested minimum timeout: 30s.
95
96 monitor (Slave role)
97 Performs a detailed status check. Suggested minimum timeout: 20s.
98 Suggested interval: 29s.
99
100 monitor (Master role)
101 Performs a detailed status check. Suggested minimum timeout: 20s.
102 Suggested interval: 60s.
103
104 meta-data
105 Retrieves resource agent metadata (internal use only). Suggested
106 minimum timeout: 5s.
107
108 validate-all
109 Performs a validation of the resource configuration. Suggested
110 minimum timeout: 30s.
111
113 The following is an example configuration for a sg_persist resource
114 using the crm(8) shell:
115
116 primitive p_sg_persist ocf:heartbeat:sg_persist \
117 params \
118 devs=string \
119 op monitor depth="0" timeout="20s" interval="29s" role="Slave" \
120 op monitor depth="0" timeout="20s" interval="60s" role="Master"
121
122 ms ms_sg_persist p_sg_persist \
123 meta notify="true" interleave="true"
124
126 The following is an example configuration for a sg_persist resource
127 using pcs(8)
128
129 pcs resource create p_sg_persist ocf:heartbeat:sg_persist \
130 devs=string \
131 op monitor OCF_CHECK_LEVEL="0" timeout="20s" interval="29s" role="Slave" \
132 op monitor OCF_CHECK_LEVEL="0" timeout="20s" interval="60s" role="Master" promotable
133
135 http://clusterlabs.org/
136
138 ClusterLabs contributors (see the resource agent source for information
139 about individual authors)
140
141
142
143resource-agents UNKNOWN 03/25/2021 OCF_HEARTBEAT_SG_PER(7)