113 lines · plain
1.. SPDX-License-Identifier: GPL-2.02 3================4uevents and GFS25================6 7During the lifetime of a GFS2 mount, a number of uevents are generated.8This document explains what the events are and what they are used9for (by gfs_controld in gfs2-utils).10 11A list of GFS2 uevents12======================13 141. ADD15------16 17The ADD event occurs at mount time. It will always be the first18uevent generated by the newly created filesystem. If the mount19is successful, an ONLINE uevent will follow. If it is not successful20then a REMOVE uevent will follow.21 22The ADD uevent has two environment variables: SPECTATOR=[0|1]23and RDONLY=[0|1] that specify the spectator status (a read-only mount24with no journal assigned), and read-only (with journal assigned) status25of the filesystem respectively.26 272. ONLINE28---------29 30The ONLINE uevent is generated after a successful mount or remount. It31has the same environment variables as the ADD uevent. The ONLINE32uevent, along with the two environment variables for spectator and33RDONLY are a relatively recent addition (2.6.32-rc+) and will not34be generated by older kernels.35 363. CHANGE37---------38 39The CHANGE uevent is used in two places. One is when reporting the40successful mount of the filesystem by the first node (FIRSTMOUNT=Done).41This is used as a signal by gfs_controld that it is then ok for other42nodes in the cluster to mount the filesystem.43 44The other CHANGE uevent is used to inform of the completion45of journal recovery for one of the filesystems journals. It has46two environment variables, JID= which specifies the journal id which47has just been recovered, and RECOVERY=[Done|Failed] to indicate the48success (or otherwise) of the operation. These uevents are generated49for every journal recovered, whether it is during the initial mount50process or as the result of gfs_controld requesting a specific journal51recovery via the /sys/fs/gfs2/<fsname>/lock_module/recovery file.52 53Because the CHANGE uevent was used (in early versions of gfs_controld)54without checking the environment variables to discover the state, we55cannot add any more functions to it without running the risk of56someone using an older version of the user tools and breaking their57cluster. For this reason the ONLINE uevent was used when adding a new58uevent for a successful mount or remount.59 604. OFFLINE61----------62 63The OFFLINE uevent is only generated due to filesystem errors and is used64as part of the "withdraw" mechanism. Currently this doesn't give any65information about what the error is, which is something that needs to66be fixed.67 685. REMOVE69---------70 71The REMOVE uevent is generated at the end of an unsuccessful mount72or at the end of a umount of the filesystem. All REMOVE uevents will73have been preceded by at least an ADD uevent for the same filesystem,74and unlike the other uevents is generated automatically by the kernel's75kobject subsystem.76 77 78Information common to all GFS2 uevents (uevent environment variables)79=====================================================================80 811. LOCKTABLE=82--------------83 84The LOCKTABLE is a string, as supplied on the mount command85line (locktable=) or via fstab. It is used as a filesystem label86as well as providing the information for a lock_dlm mount to be87able to join the cluster.88 892. LOCKPROTO=90-------------91 92The LOCKPROTO is a string, and its value depends on what is set93on the mount command line, or via fstab. It will be either94lock_nolock or lock_dlm. In the future other lock managers95may be supported.96 973. JOURNALID=98-------------99 100If a journal is in use by the filesystem (journals are not101assigned for spectator mounts) then this will give the102numeric journal id in all GFS2 uevents.103 1044. UUID=105--------106 107With recent versions of gfs2-utils, mkfs.gfs2 writes a UUID108into the filesystem superblock. If it exists, this will109be included in every uevent relating to the filesystem.110 111 112 113