1OCF_HEARTBEAT_MDRAID(7) OCF resource agents OCF_HEARTBEAT_MDRAID(7)
2
6 ocf_heartbeat_mdraid - Manages Linux software RAID (MD) devices on
7 shared storage
8
10 mdraid [start | stop | monitor | meta-data | validate-all]
11
13 This resource agent manages Linux software RAID (MD) devices on a
14 shared storage medium ensuring that non-clustered MD arrays are
15 prohibited from starting cloned (which would cause data corruption
16 (e.g., on raid6 arrays) unless forced (see force_clones parameter).
17 Clustered MD RAID layouts (see below) will be discovered and allowed
18 cloning by default; no need to set force_clones.
19 It uses mdadm(8) to start, stop, and monitor the MD devices.
21 Supported clustered (i.e., clonable active-active) arrays are linear,
23 raid0, and clustered raid1/raid10 (i.e. mdadm(8) created with
24 --bitmap=clustered).
26
28 mdadm_conf
29 The MD RAID configuration file (e.g., /etc/mdadm.conf).
30 (required, string, no default)
32 md_dev
34 MD array block device to use (e.g., /dev/md0 or /dev/md/3). With
35 shared access to the array's storage, this should preferably be a
36 clustered raid1 or raid10 array created with --bitmap=clustered,
37 assuming its resource will be cloned (i.e., active-active access).
38 Be sure to disable auto-assembly for the resource-managed arrays!
40 (required, string, no default)
42 force_stop
44 If processes or kernel threads are using the array, it cannot be
45 stopped. We will try to stop processes, first by sending TERM and
46 then, if that doesn't help in seconds, using KILL. The lsof(8)
47 program is required to get the list of array users. Of course, the
48 kernel threads cannot be stopped this way. If the processes are
49 critical for data integrity, then set this parameter to false. Note
50 that in that case the stop operation will fail and the node will be
51 fenced.
52 (optional, boolean, default false)
54 wait_for_udev
56 Wait until udevd creates a device in the start operation. On a
57 normally loaded host this should happen quickly, but you may be
58 unlucky. If you are not using udev set this to "no".
59 (optional, boolean, default true)
61 force_clones
63 Activating the same, non-clustered MD RAID array (i.e. single-host
64 raid1/4/5/6/10) on multiple nodes at the same time will result in
65 data corruption and thus is forbidden by default.
66 A safe example could be an (exotic) array that is only named
68 identically across all nodes, but is in fact based on distinct
69 (non-shared) storage.
70 Only set this to "true" if you know what you are doing!
72 (optional, boolean, default false)
74
76 This resource agent supports the following actions (operations):
77 start
79 Starts the resource. Suggested minimum timeout: 20s.
80 stop
82 Stops the resource. Suggested minimum timeout: 20s.
83 monitor
85 Performs a detailed status check. Suggested minimum timeout: 20s.
86 Suggested interval: 10s.
87 validate-all
89 Performs a validation of the resource configuration. Suggested
90 minimum timeout: 5s.
91 meta-data
93 Retrieves resource agent metadata (internal use only). Suggested
94 minimum timeout: 5s.
95
97 The following is an example configuration for a mdraid resource using
98 the crm(8) shell:
99 primitive p_mdraid ocf:heartbeat:mdraid \
101 params \
102 mdadm_conf=string \
103 md_dev=string \
104 op monitor depth="0" timeout="20s" interval="10s"
105
107 The following is an example configuration for a mdraid resource using
108 pcs(8)
109 pcs resource create p_mdraid ocf:heartbeat:mdraid \
111 mdadm_conf=string \
112 md_dev=string \
113 op monitor OCF_CHECK_LEVEL="0" timeout="20s" interval="10s"
114
116 http://clusterlabs.org/
117
119 ClusterLabs contributors (see the resource agent source for information
120 about individual authors)
121resource-agents UNKNOWN 03/25/2021 OCF_HEARTBEAT_MDRAID(7)