b=21681 Quiet bogus previously committed transno error

[fs/lustre-release.git] / lustre / ptlrpc / import.c

diff --git a/lustre/ptlrpc/import.c b/lustre/ptlrpc/import.c
index d5a1e74..e50a11d 100644 (file)
--- a/lustre/ptlrpc/import.c
+++ b/lustre/ptlrpc/import.c
@@ -667,6 +667,7 @@ int ptlrpc_connect_import(struct obd_import *imp, char *new_uuid)
          * the server is updated on-the-fly we will get the new features. */
         imp->imp_connect_data.ocd_connect_flags = imp->imp_connect_flags_orig;
         imp->imp_msghdr_flags &= ~MSGHDR_AT_SUPPORT;
+        imp->imp_msghdr_flags &= ~MSGHDR_CKSUM_INCOMPAT18;
 
         rc = obd_reconnect(NULL, imp->imp_obd->obd_self_export, obd,
                            &obd->obd_uuid, &imp->imp_connect_data, NULL);
@@ -943,7 +944,8 @@ static int ptlrpc_connect_interpret(const struct lu_env *env,
                        "after reconnect. We should LBUG right here.\n");
         }
 
-        if (lustre_msg_get_last_committed(request->rq_repmsg) <
+        if (lustre_msg_get_last_committed(request->rq_repmsg) > 0 &&
+            lustre_msg_get_last_committed(request->rq_repmsg) <
             aa->pcaa_peer_committed) {
                 CERROR("%s went back in time (transno "LPD64
                        " was previously committed, server now claims "LPD64
@@ -1112,6 +1114,12 @@ finish:
                 else
                         imp->imp_msghdr_flags &= ~MSGHDR_AT_SUPPORT;
 
+                if ((ocd->ocd_connect_flags & OBD_CONNECT_FULL20) &&
+                    (imp->imp_msg_magic == LUSTRE_MSG_MAGIC_V2))
+                        imp->imp_msghdr_flags |= MSGHDR_CKSUM_INCOMPAT18;
+                else
+                        imp->imp_msghdr_flags &= ~MSGHDR_CKSUM_INCOMPAT18;
+
                 LASSERT((cli->cl_max_pages_per_rpc <= PTLRPC_MAX_BRW_PAGES) &&
                         (cli->cl_max_pages_per_rpc > 0));
         }
@@ -1197,9 +1205,6 @@ static int completed_replay_interpret(const struct lu_env *env,
                         CDEBUG(D_WARNING,
                                "%s: version recovery fails, reconnecting\n",
                                req->rq_import->imp_obd->obd_name);
-                        cfs_spin_lock(&req->rq_import->imp_lock);
-                        req->rq_import->imp_vbr_failed = 0;
-                        cfs_spin_unlock(&req->rq_import->imp_lock);
                 } else {
                         CDEBUG(D_HA, "%s: LAST_REPLAY message error: %d, "
                                      "reconnecting\n",
@@ -1289,6 +1294,10 @@ int ptlrpc_import_recovery_state_machine(struct obd_import *imp)
                 CDEBUG(D_HA, "evicted from %s@%s; invalidating\n",
                        obd2cli_tgt(imp->imp_obd),
                        imp->imp_connection->c_remote_uuid.uuid);
+                /* reset vbr_failed flag upon eviction */
+                cfs_spin_lock(&imp->imp_lock);
+                imp->imp_vbr_failed = 0;
+                cfs_spin_unlock(&imp->imp_lock);
 
 #ifdef __KERNEL__
                 /* bug 17802:  XXX client_disconnect_export vs connect request
@@ -1370,11 +1379,6 @@ out:
         RETURN(rc);
 }
 
-static int back_to_sleep(void *unused)
-{
-        return 0;
-}
-
 int ptlrpc_disconnect_import(struct obd_import *imp, int noclose)
 {
         struct ptlrpc_request *req;