1 This directory is where you should put any local or application
2 specific event scripts for ctdb to call.
4 All event scripts start with the prefic 'NN.' where N is a digit.
5 The event scripts are run in sequence based on NN.
6 Thus 10.interfaces will be run before 60.nfs.
8 Each NN must be unique and duplicates will cause undefined behaviour.
9 I.e. having both 10.interfaces and 10.otherstuff is not allowed.
12 As a special case, any eventscript that ends with a '~' character will be
13 ignored since this is a common postfix that some editors will append to
14 older versions of a file.
17 The eventscripts are called with varying number of arguments.
18 The first argument is the "event" and the rest of the arguments depend
19 on which event was triggered.
21 The events currently implemented are
23 This event does not take any additional arguments.
24 This event is only invoked once, when ctdb is starting up.
25 This event is used to wait for the service to start and all
26 resources for the service becoming available.
28 This is used to prevent ctdb from starting up and advertize its
29 services until all dependent services have become available.
31 All services that are managed by ctdb should implement this
32 event and use it to start the service.
34 Example: 50.samba uses this event to start the samba daemon
35 and then wait until samba and all its associated services have
36 become available. It then also proceeds to wait until all
37 shares have become available.
40 This event is called when the ctdb service is shuting down.
42 All services that are managed by ctdb should implement this event
43 and use it to perform a controlled shutdown of the service.
45 Example: 60.nfs uses this event to shut down nfs and all associated
46 services and stop exporting any shares when this event is invoked.
49 This event is invoked every X number of seconds.
50 The interval can be configured using the MonitorInterval tunable
51 but defaults to 15 seconds.
53 This event is triggered by ctdb to continously monitor that all
54 managed services are healthy.
55 When invoked, the event script will check that the service is healthy
56 and return 0 if so. If the service is not healthy the event script
57 should return non zero.
59 If a service returns nonzero from this script this will cause ctdb
60 to consider the node status as UNHEALTHY and will cause the public
61 address and all associated services to be failed over to a different
64 All managed services should implement this event.
66 Example: 10.interfaces which checks that the public interface (if used)
67 is healthy, i.e. it has a physical link established.
70 This event is triggered everytime the node takes over a public ip
71 address during recovery.
72 This event takes three additional arguments :
73 'interface' 'ipaddress' and 'netmask'
75 This event will always be followed by a 'recovered' event onse
76 all ipaddresses have been reassigned to new nodes and the ctdb database
78 If multiple ip addresses are reassigned during recovery it is
79 possible to get several 'takeip' events followed by a single
82 Since there might involve substantial work for the service when an ip
83 address is taken over and since multiple ip addresses might be taken
84 over in a single recovery it is often best to only mark which addresses
85 are being taken over in this event and defer the actual work to
86 reconfigure or restart the services until the 'recovered' event.
88 Example: 60.nfs which just records which ip addresses are being taken
89 over into a local state directory and which defers the actual
90 restart of the services until the 'recovered' event.
94 This event is triggered everytime the node releases a public ip
95 address during recovery.
96 This event takes three additional arguments :
97 'interface' 'ipaddress' and 'netmask'
99 In all other regards this event is analog to the 'takeip' event above.
104 This event is triggered everytime a full ctdb recovery has completed
105 and all public ip addresses have been reassigned among the nodes.
107 Example: 60.nfs which if the ip address configuration has changed
108 during the recovery (i.e. if addresses have been taken over or
109 released) will kill off any tcp connections that exist for that
110 service and also send out statd notifications to all registered
114 Additional note for takeip, releaseip, recovered:
116 ALL services that depend on the ip address configuration of the node must
117 implement all three of these events.
119 ALL services that use TCP should also implement these events and at least
120 kill off any tcp connections to the service if the ip address config has
121 changed in a similar fashion to how 60.nfs does it.
122 The reason one must do this is that ESTABLISHED tcp connections may survive
123 when an ip address is released and removed from the host until the ip address
125 Any tcp connections that survive a release/takeip sequence can potentially
126 cause the client/server tcp connection to get out of sync with sequence and
127 ack numbers and cause a disruptive ack storm.