drbd/0020-drbd-do-not-allow-auto-demote-to-be-interrupted-by-s.patch

67 lines
2.3 KiB
Diff

From c8e2a3c4355b4794267cd6e58a074802b4607cb9 Mon Sep 17 00:00:00 2001
From: Joel Colledge <joel.colledge@linbit.com>
Date: Fri, 22 Sep 2023 16:57:24 +0200
Subject: [PATCH 20/20] drbd: do not allow auto-demote to be interrupted by
signal
Pending signals can mess up auto-demote:
drbd res: Preparing cluster-wide state change 671410162 (0->-1 3/2)
drbd res: Aborting cluster-wide state change 671410162 (6ms) rv = -21
drbd res: Auto-demote failed: Interrupted state change
After this state change failure no process has the DRBD device open, but
DRBD remains Primary.
---
drbd/drbd_main.c | 33 +++++++++++++++++++++++----------
1 file changed, 23 insertions(+), 10 deletions(-)
diff --git a/drbd/drbd_main.c b/drbd/drbd_main.c
index 1864861db21d..0719229f210e 100644
--- a/drbd/drbd_main.c
+++ b/drbd/drbd_main.c
@@ -2851,17 +2851,30 @@ static void drbd_release(struct gendisk *gd)
!test_and_set_bit(DESTROYING_DEV, &device->flags))
call_rcu(&device->rcu, drbd_reclaim_device);
- if (resource->res_opts.auto_promote) {
- enum drbd_state_rv rv;
+ if (resource->res_opts.auto_promote &&
+ open_rw_cnt == 0 &&
+ resource->role[NOW] == R_PRIMARY &&
+ !test_bit(EXPLICIT_PRIMARY, &resource->flags)) {
+ sigset_t mask, oldmask;
+ int rv;
+
+ /*
+ * Auto-demote is triggered by the last opener releasing the
+ * DRBD device. However, it is an implicit action, so it should
+ * not be affected by the state of the process. In particular,
+ * it should ignore any pending signals. It may be the case
+ * that the process is releasing DRBD because it is being
+ * terminated using a signal.
+ */
+ sigfillset(&mask);
+ sigprocmask(SIG_BLOCK, &mask, &oldmask);
- if (was_writable && open_rw_cnt == 0 &&
- resource->role[NOW] == R_PRIMARY &&
- !test_bit(EXPLICIT_PRIMARY, &resource->flags)) {
- rv = drbd_set_role(resource, R_SECONDARY, false, "auto-demote", NULL);
- if (rv < SS_SUCCESS)
- drbd_warn(resource, "Auto-demote failed: %s (%d)\n",
- drbd_set_st_err_str(rv), rv);
- }
+ rv = drbd_set_role(resource, R_SECONDARY, false, "auto-demote", NULL);
+ if (rv < SS_SUCCESS)
+ drbd_warn(resource, "Auto-demote failed: %s (%d)\n",
+ drbd_set_st_err_str(rv), rv);
+
+ sigprocmask(SIG_SETMASK, &oldmask, NULL);
}
if (open_ro_cnt == 0 && open_rw_cnt == 0 && resource->fail_io[NOW]) {
--
2.35.3