87bc6c8acd
- Detect an unresponsive subdaemon - Handle certain probe failures as stopped instead of failed - Update pcmk_delay_base option meta-data - Avoid crash when using clone notifications - Retry Corosync shutdown tracking if first attempt fails - Improve display of failed actions - Resolves: rhbz1707851 - Resolves: rhbz2039982 - Resolves: rhbz2032032 - Resolves: rhbz2040443 - Resolves: rhbz2042367 - Resolves: rhbz2042546
30 lines
1.1 KiB
Diff
30 lines
1.1 KiB
Diff
From e8bf0161b872267f1bb7143a9866fdc15ec218f2 Mon Sep 17 00:00:00 2001
|
|
From: Jan Friesse <jfriesse@redhat.com>
|
|
Date: Tue, 18 Jan 2022 16:35:24 +0100
|
|
Subject: [PATCH] Fix: corosync: Repeat corosync_cfg_trackstart
|
|
|
|
corosync_cfg_trackstart can fail with CS_ERR_TRY_AGAIN failure so
|
|
(similarly as for corosync_cfg_local_get, ...) handle failure with
|
|
using cs_repeat macro.
|
|
---
|
|
daemons/pacemakerd/pcmkd_corosync.c | 3 ++-
|
|
1 file changed, 2 insertions(+), 1 deletion(-)
|
|
|
|
diff --git a/daemons/pacemakerd/pcmkd_corosync.c b/daemons/pacemakerd/pcmkd_corosync.c
|
|
index 7990bc43c5..cd7a40321d 100644
|
|
--- a/daemons/pacemakerd/pcmkd_corosync.c
|
|
+++ b/daemons/pacemakerd/pcmkd_corosync.c
|
|
@@ -186,7 +186,8 @@ cluster_connect_cfg(void)
|
|
crm_debug("Corosync reports local node ID is %lu", (unsigned long) nodeid);
|
|
|
|
#ifdef HAVE_COROSYNC_CFG_TRACKSTART
|
|
- rc = corosync_cfg_trackstart(cfg_handle, 0);
|
|
+ retries = 0;
|
|
+ cs_repeat(retries, 30, rc = corosync_cfg_trackstart(cfg_handle, 0));
|
|
if (rc != CS_OK) {
|
|
crm_crit("Could not enable Corosync CFG shutdown tracker: %s " CRM_XS " rc=%d",
|
|
cs_strerror(rc), rc);
|
|
--
|
|
2.27.0
|
|
|