Initial support for mdadm RAID arrays

This patch adds initial support for handling MD RAID arrays.  Adding
a pass at the very end of the shutdown sequence to signal any external
RAID controller to clear its state before we proceed to halt/reboot.

Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
This commit is contained in:
Joachim Nilsson committed 2016-08-14 21:26:18 +02:00
1 parent f4327910d9
commit 078982d3ec
4 files changed
+79 -4

No files matched your search

+3 -1
View File
@@ -86,6 +86,7 @@ static uev_t sighup_watcher, sigint_watcher, sigpwr_watcher;
static uev_t sigchld_watcher, sigsegv_watcher;
static uev_t sigstop_watcher, sigtstp_watcher, sigcont_watcher;
void mdadm_wait(void);
void unmount_tmpfs(void);
void unmount_regular(void);
@@ -155,7 +156,8 @@ void do_shutdown(shutop_t op)
run("/bin/mount -n -o remount,ro dummydev /");
run("/bin/mount -n -o remount,ro /");
/* XXX: Call mdadm to mark RAID array(s) as clean before halting. */
/* Call mdadm to mark any RAID array(s) as clean before halting. */
mdadm_wait();
_d("%s.", op == SHUT_REBOOT ? "Rebooting" : "Halting");
if (op == SHUT_REBOOT)