mdadm: Fix socket connection failure when mdmon runs in foreground mode.

While creating an IMSM RAID, mdadm will wait for the mdmon main process
to finish if mdmon runs in forking mode. This is because with
"Type=forking" in the mdmon service unit file, "systemctl start service"
will block until the main process of mdmon exits. At that moment, mdmon
has already created the socket, so the subsequent socket connect from
mdadm will succeed.

However, when mdmon runs in foreground mode (without "Type=forking" in
the service unit file), "systemctl start service" will return once the
mdmon process starts. This causes mdadm and mdmon to run in parallel,
which may lead to a socket connection failure since mdmon has not yet
initialized the socket when mdadm tries to connect. If the next
instruction/command is to access this device and try to  write to it, a
permission error will occur since mdmon has not yet set the array to RW
mode.

Signed-off-by: Shminderjit Singh <shminderjit.singh@oracle.com>

diff --git a/Create.c b/Create.c
index bd4875e4a408b76a6fe264665140fde4b3804e6c..479c27158a48c9e74d98b923603ae6e25f9b605a 100644 (file)
--- a/Create.c
+++ b/Create.c
@@ -1344,9 +1344,11 @@ int Create(struct supertype *st, struct mddev_ident *ident, int subdevs,
                if (c->verbose >= 0)
                        pr_info("array %s started.\n", chosen_name);
                if (st->ss->external && st->container_devnm[0]) {
-                       if (need_mdmon)
+                       if (need_mdmon) {
                                start_mdmon(st->container_devnm);
-
+                               if (wait_for_mdmon_control_socket(st->container_devnm) != MDADM_STATUS_SUCCESS)
+                                       goto abort;
+                       }
                        ping_monitor(st->container_devnm);
                        close(container_fd);
                }

diff --git a/mdadm.h b/mdadm.h
index e9f764a23b1f0dc4bcaa1b7e56ec33acc9eafab5..27009154297f38140fbdffe3826551ce36e8d4ef 100644 (file)
--- a/mdadm.h
+++ b/mdadm.h
@@ -1776,6 +1776,7 @@ extern int is_subarray_active(char *subarray, char *devname);
 extern int open_subarray(char *dev, char *subarray, struct supertype *st, int quiet);
 extern struct superswitch *version_to_superswitch(char *vers);
 
+extern mdadm_status_t wait_for_mdmon_control_socket(const char *container_devnm);
 extern int mdmon_running(const char *devnm);
 extern int mdmon_pid(const char *devnm);
 extern mdadm_status_t wait_for_mdmon(const char *devnm);

diff --git a/util.c b/util.c
index 48c97545a42a619235eaf3125a26ce133d53479b..908f84306ed2f543b68f9eda8107fdfcb40179d4 100644 (file)
--- a/util.c
+++ b/util.c
@@ -1932,6 +1932,51 @@ int mdmon_running(const char *devnm)
        return 0;
 }
 
+/*
+ * wait_for_mdmon_control_socket() - Waits for mdmon control socket
+ * to be created within specified time.
+ * @container_devnm: Device for which mdmon control socket should start.
+ *
+ * In foreground mode, when mdadm is trying to connect to control
+ * socket it is possible that the mdmon has not created it yet.
+ * Give some time to mdmon to create socket. Timeout set to 2 sec.
+ *
+ * Return: MDADM_STATUS_SUCCESS if connect succeed, otherwise return
+ * error code.
+ */
+mdadm_status_t wait_for_mdmon_control_socket(const char *container_devnm)
+{
+       enum mdadm_status status = MDADM_STATUS_SUCCESS;
+       int sfd, rv, retry_count = 0;
+       struct sockaddr_un addr;
+       char path[PATH_MAX];
+
+       snprintf(path, PATH_MAX, "%s/%s.sock", MDMON_DIR, container_devnm);
+       sfd = socket(PF_LOCAL, SOCK_STREAM, 0);
+       if (!is_fd_valid(sfd))
+               return MDADM_STATUS_ERROR;
+
+       addr.sun_family = PF_LOCAL;
+        strncpy(addr.sun_path, path, sizeof(addr.sun_path) - 1);
+        addr.sun_path[sizeof(addr.sun_path) - 1] = '\0';
+
+       for (retry_count = 0; retry_count < 10; retry_count++) {
+               rv = connect(sfd, (struct sockaddr*)&addr, sizeof(addr));
+               if (rv < 0) {
+                       sleep_for(0, MSEC_TO_NSEC(200), true);
+                       continue;
+               }
+               break;
+       }
+
+       if (rv < 0) {
+               pr_err("Failed to connect to control socket.\n");
+               status = MDADM_STATUS_ERROR;
+       }
+       close(sfd);
+       return status;
+}
+
 /*
  * wait_for_mdmon() - Waits for mdmon within specified time.
  * @devnm: Device for which mdmon should start.