Exit offline mode only if server is available.

This patch adds periodic check to test if
backend can exit offline mode and only
marks backend as not offline if server
for the service is available.

Prior to this patch we marked backend as not
offline if the offline_timeout was reached
without checking for the server availability
and when the next request failed again
we switched back to the offline mode.
This caused significant slowdowns in some
edge cases.

Fixes:
https://fedorahosted.org/sssd/ticket/2355

Reviewed-by: Jakub Hrozek <jhrozek@redhat.com>

2 files changed, 43 insertions, 7 deletions

diff --git a/src/providers/data_provider_be.c b/src/providers/data_provider_be.c
index e5b13b967..1ba32dfb7 100644
--- a/src/providers/data_provider_be.c
+++ b/src/providers/data_provider_be.c
@@ -450,24 +450,57 @@ static void be_queue_next_request(struct be_req *be_req, enum bet_type type)
 
 bool be_is_offline(struct be_ctx *ctx)
 {
-    time_t now = time(NULL);
+    return ctx->offstat.offline;
+}
 
-    /* check if we are past the offline blackout timeout */
-    /* FIXME: get offline_timeout from configuration */
-    if (ctx->offstat.went_offline + 60 < now) {
-        ctx->offstat.offline = false;
-    }
+static void check_if_online(struct be_ctx *ctx);
 
-    return ctx->offstat.offline;
+static errno_t
+try_to_go_online(TALLOC_CTX *mem_ctx,
+                 struct tevent_context *ev,
+                 struct be_ctx *be_ctx,
+                 struct be_ptask *be_ptask,
+                 void *be_ctx_void)
+{
+    struct be_ctx *ctx = (struct be_ctx*) be_ctx_void;
+
+    check_if_online(ctx);
+    return EOK;
 }
 
 void be_mark_offline(struct be_ctx *ctx)
 {
+    int offline_timeout = 60;
+    errno_t ret;
+
     DEBUG(SSSDBG_TRACE_INTERNAL, "Going offline!\n");
 
     ctx->offstat.went_offline = time(NULL);
     ctx->offstat.offline = true;
     ctx->run_online_cb = true;
+
+    if (ctx->check_if_online_ptask == NULL) {
+        /* This is the first time we go offline - create a periodic task
+         * to check if we can switch to online. */
+        ret = be_ptask_create_sync(ctx, ctx,
+                                   offline_timeout, offline_timeout,
+                                   offline_timeout, 30, offline_timeout,
+                                   BE_PTASK_OFFLINE_EXECUTE,
+                                   3600 /* max_backoff */,
+                                   try_to_go_online,
+                                   ctx, "Check if online (periodic)",
+                                   &ctx->check_if_online_ptask);
+        if (ret != EOK) {
+            DEBUG(SSSDBG_FATAL_FAILURE,
+                  "be_ptask_create_sync failed [%d]: %s\n",
+                  ret, sss_strerror(ret));
+        }
+    } else {
+        /* Periodic task was already created. Just enable it. */
+        DEBUG(SSSDBG_TRACE_INTERNAL, "Enable check_if_online_ptask.\n");
+        be_ptask_enable(ctx->check_if_online_ptask);
+    }
+
     be_run_offline_cb(ctx);
 }
 
@@ -475,6 +508,7 @@ static void be_reset_offline(struct be_ctx *ctx)
 {
     ctx->offstat.went_offline = 0;
     ctx->offstat.offline = false;
+    be_ptask_disable(ctx->check_if_online_ptask);
     be_run_online_cb(ctx);
 }
 
diff --git a/src/providers/dp_backend.h b/src/providers/dp_backend.h
index 8e3a68a5b..075681ff9 100644
--- a/src/providers/dp_backend.h
+++ b/src/providers/dp_backend.h
@@ -133,6 +133,8 @@ struct be_ctx {
     struct be_cb *unconditional_online_cb_list;
 
     struct be_offline_status offstat;
+    /* Periodicly check if we can go online. */
+    struct be_ptask *check_if_online_ptask;
 
     struct sbus_connection *mon_conn;
     struct sbus_connection *sbus_srv;