blob: fd966dc9979a9c37346633b051beaaf9686604ba [file] [log] [blame]
Steven Whitehouse0aa87442009-08-14 15:24:46 +01001 uevents and GFS2
2 ==================
3
4During the lifetime of a GFS2 mount, a number of uevents are generated.
5This document explains what the events are and what they are used
6for (by gfs_controld in gfs2-utils).
7
8A list of GFS2 uevents
9-----------------------
10
111. ADD
12
13The ADD event occurs at mount time. It will always be the first
14uevent generated by the newly created filesystem. If the mount
15is successful, an ONLINE uevent will follow. If it is not successful
16then a REMOVE uevent will follow.
17
18The ADD uevent has two environment variables: SPECTATOR=[0|1]
19and RDONLY=[0|1] that specify the spectator status (a read-only mount
20with no journal assigned), and read-only (with journal assigned) status
21of the filesystem respectively.
22
232. ONLINE
24
25The ONLINE uevent is generated after a successful mount or remount. It
26has the same environment variables as the ADD uevent. The ONLINE
27uevent, along with the two environment variables for spectator and
28RDONLY are a relatively recent addition (2.6.32-rc+) and will not
29be generated by older kernels.
30
313. CHANGE
32
33The CHANGE uevent is used in two places. One is when reporting the
34successful mount of the filesystem by the first node (FIRSTMOUNT=Done).
35This is used as a signal by gfs_controld that it is then ok for other
36nodes in the cluster to mount the filesystem.
37
38The other CHANGE uevent is used to inform of the completion
39of journal recovery for one of the filesystems journals. It has
40two environment variables, JID= which specifies the journal id which
41has just been recovered, and RECOVERY=[Done|Failed] to indicate the
42success (or otherwise) of the operation. These uevents are generated
43for every journal recovered, whether it is during the initial mount
44process or as the result of gfs_controld requesting a specific journal
45recovery via the /sys/fs/gfs2/<fsname>/lock_module/recovery file.
46
47Because the CHANGE uevent was used (in early versions of gfs_controld)
48without checking the environment variables to discover the state, we
49cannot add any more functions to it without running the risk of
50someone using an older version of the user tools and breaking their
51cluster. For this reason the ONLINE uevent was used when adding a new
52uevent for a successful mount or remount.
53
544. OFFLINE
55
56The OFFLINE uevent is only generated due to filesystem errors and is used
57as part of the "withdraw" mechanism. Currently this doesn't give any
58information about what the error is, which is something that needs to
59be fixed.
60
615. REMOVE
62
63The REMOVE uevent is generated at the end of an unsuccessful mount
64or at the end of a umount of the filesystem. All REMOVE uevents will
65have been preceeded by at least an ADD uevent for the same fileystem,
66and unlike the other uevents is generated automatically by the kernel's
67kobject subsystem.
68
69
70Information common to all GFS2 uevents (uevent environment variables)
71----------------------------------------------------------------------
72
731. LOCKTABLE=
74
75The LOCKTABLE is a string, as supplied on the mount command
76line (locktable=) or via fstab. It is used as a filesystem label
77as well as providing the information for a lock_dlm mount to be
78able to join the cluster.
79
802. LOCKPROTO=
81
82The LOCKPROTO is a string, and its value depends on what is set
83on the mount command line, or via fstab. It will be either
84lock_nolock or lock_dlm. In the future other lock managers
85may be supported.
86
873. JOURNALID=
88
89If a journal is in use by the filesystem (journals are not
90assigned for spectator mounts) then this will give the
91numeric journal id in all GFS2 uevents.
92
934. UUID=
94
95With recent versions of gfs2-utils, mkfs.gfs2 writes a UUID
96into the filesystem superblock. If it exists, this will
97be included in every uevent relating to the filesystem.
98
99
100