1o2cb(7)                       OCFS2 Manual Pages                       o2cb(7)
2
3
4

NAME

6       o2cb - Default cluster stack for the OCFS2 file system.
7

DESCRIPTION

9       o2cb  is  the  default  cluster  stack  for  the  OCFS2 file system. It
10       includes a node manager (o2nm) to keep track of the nodes in the  clus‐
11       ter,  a  heartbeat  agent  (o2hb) to detect live nodes, a network agent
12       (o2net) for intra-cluster node communication  and  a  distributed  lock
13       manager  (o2dlm)  to keep track of lock resources. All these components
14       are in-kernel. It also includes an in-memory  file  system,  dlmfs,  to
15       allow userspace to access the in-kernel dlm.
16
17       This    cluster    stack   has   two   configuration   files,   namely,
18       /etc/ocfs2/cluster.conf and  /etc/sysconfig/o2cb.  Whereas  the  former
19       keeps  track of the cluster layout, the latter keeps track of the clus‐
20       ter timeouts. Both files are only read  when  the  cluster  is  brought
21       online.  Values  in  use  by  the  online cluster can be perused in the
22       /sys/kernel/config/cluster directory structure.
23
24

CONFIGURATION

26       The cluster layout is specified in /etc/ocfs2/cluster.conf. While it is
27       easier   to  populate  and  propagate  this  configuration  file  using
28       ocfs2console(8), one can also do it by manually  as  long  as  care  is
29       taken to format the file correctly.
30
31       While  the console utility is intuitive to use, there are few points to
32       keep in mind.
33
34            1. The node name needs to match the hostname. It does not need  to
35       include  the  domain  name.  For  example,  appserver.oracle.com can be
36       appserver.
37
38            2. The IP address need not be the one associated with  that  host‐
39       name.  As  in, any valid IP address on that node can be used. O2CB will
40       not attempt to match the node name (hostname)  with  the  specified  IP
41       address.
42
43       For  best performance, use of a private interconnect (lower latency) is
44       recommended.
45
46       The cluster.conf file is in a stanza format with two types of  stanzas,
47       namely,  cluster and node. A typical cluster.conf will have one cluster
48       stanza and multiple node stanzas.
49
50       The cluster stanza has two parameters:
51
52       node_count
53              Total number of nodes in the cluster
54
55       name   Name of the cluster
56
57
58       The node stanza has five parameters:
59
60       ip_port
61              IP port
62
63       ip_address
64              IP address
65
66       number Unique node number from 0-254
67
68       name   Hostname
69
70       cluster
71              Name of the cluster
72
73       Users  populating  cluster.conf  manually  should  follow  the   format
74       strictly. As in, stanza header should start at the first column and end
75       with a colon, stanza parameters should start after a tab, a blank  line
76       should demarcate each stanza and care taken to avoid stray whitespaces.
77
78       The  O2CB cluster timeouts are specified in /etc/sysconfig/o2cb and can
79       be configured using the o2cb init script.
80
81       These timeouts are used by the O2CB clusterstack to determine whether a
82       node  is dead or alive. While the use of default values is recommended,
83       users can experiment with other values if the defaults are causing spu‐
84       rious fencing.
85
86       The cluster timeouts are:
87
88       Heartbeat Dead Threshold
89              The  Disk  Heartbeat  timeout is the number of two second itera‐
90              tions before a node is considered dead. The exact  formula  used
91              to convert the timeout in seconds to the number of iterations is
92              as follows:
93
94              O2CB_HEARTBEAT_THRESHOLD = (((timeout in seconds) / 2) + 1)
95
96              For e.g., to specify a 60 sec timeout, set it  to  31.  For  120
97              secs,  set  it  to  61.  The default for this timeout is 60 secs
98              (O2CB_HEARTBEAT_THRESHOLD = 31).
99
100
101       Network Idle Timeout
102              The Network Idle timeout  specifies  the  time  in  milliseconds
103              before  a  network connection is considered dead. It defaults to
104              30000 ms.
105
106
107       Network Keepalive Delay
108              The Network Keepalive specifies the maximum delay  in  millisec‐
109              onds  before a keepalive packet is sent to another node to check
110              whether it is alive or not.  If  the  node  is  alive,  it  will
111              respond. Its defaults to 2000 ms.
112
113
114       Network Reconnect Delay
115              The  Network  Reconnect specifies the minimum delay in millisec‐
116              onds between connection attempts. It defaults to 2000 ms.
117

EXAMPLES

119       A sample /etc/ocfs2/cluster.conf.
120
121       cluster:
122           node_count = 3
123           name = webcluster
124
125       node:
126           ip_port = 7777
127           ip_address = 192.168.0.107
128           number = 7
129           name = node7
130           cluster = webcluster
131
132       node:
133           ip_port = 7777
134           ip_address = 192.168.0.106
135           number = 6
136           name = node6
137           cluster = webcluster
138
139       node:
140           ip_port = 7777
141           ip_address = 192.168.0.110
142           number = 10
143           name = node10
144           cluster = webcluster
145
146

SEE ALSO

148       mkfs.ocfs2(8) fsck.ocfs2(8) tunefs.ocfs2(8) debugfs.ocfs2(8)  ocfs2con‐
149       sole(8)
150
151

AUTHORS

153       Oracle Corporation
154
155
157       Copyright © 2004, 2008 Oracle. All rights reserved.
158
159
160
161Version 1.4.3                      July 2008                           o2cb(7)
Impressum