LU-7524 fld: fld_clientlookup retries next target

[fs/lustre-release.git] / lustre / fld / fld_request.c

diff --git a/lustre/fld/fld_request.c b/lustre/fld/fld_request.c
index 2217475..9354922 100644 (file)
--- a/lustre/fld/fld_request.c
+++ b/lustre/fld/fld_request.c
@@ -27,7 +27,7 @@
  * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
  * Use is subject to license terms.
  *
- * Copyright (c) 2011, 2014, Intel Corporation.
+ * Copyright (c) 2011, 2015, Intel Corporation.
  */
 /*
  * This file is part of Lustre, http://www.lustre.org/
@@ -164,15 +164,8 @@ int fld_client_add_target(struct lu_client_fld *fld,
         LASSERT(name != NULL);
         LASSERT(tar->ft_srv != NULL || tar->ft_exp != NULL);
 
-        if (fld->lcf_flags != LUSTRE_FLD_INIT) {
-                CERROR("%s: Attempt to add target %s (idx "LPU64") "
-                       "on fly - skip it\n", fld->lcf_name, name,
-                       tar->ft_idx);
-                RETURN(0);
-        } else {
-                CDEBUG(D_INFO, "%s: Adding target %s (idx "
-                       LPU64")\n", fld->lcf_name, name, tar->ft_idx);
-        }
+       CDEBUG(D_INFO, "%s: Adding target %s (idx "LPU64")\n", fld->lcf_name,
+              name, tar->ft_idx);
 
         OBD_ALLOC_PTR(target);
         if (target == NULL)
@@ -304,10 +297,9 @@ int fld_client_init(struct lu_client_fld *fld,
                 RETURN(-EINVAL);
         }
 
-        fld->lcf_count = 0;
+       fld->lcf_count = 0;
        spin_lock_init(&fld->lcf_lock);
-        fld->lcf_hash = &fld_hash[hash];
-        fld->lcf_flags = LUSTRE_FLD_INIT;
+       fld->lcf_hash = &fld_hash[hash];
        INIT_LIST_HEAD(&fld->lcf_targets);
 
         cache_size = FLD_CLIENT_CACHE_SIZE /
@@ -433,7 +425,7 @@ again:
        }
 
        if (rc != 0) {
-               if (imp->imp_state != LUSTRE_IMP_CLOSED) {
+               if (imp->imp_state != LUSTRE_IMP_CLOSED && !imp->imp_deactive) {
                        /* Since LWP is not replayable, so it will keep
                         * trying unless umount happens, otherwise it would
                         * cause unecessary failure of the application. */
@@ -470,21 +462,21 @@ int fld_client_lookup(struct lu_client_fld *fld, u64 seq, u32 *mds,
 {
        struct lu_seq_range res = { 0 };
        struct lu_fld_target *target;
+       struct lu_fld_target *origin;
        int rc;
        ENTRY;
 
-        fld->lcf_flags |= LUSTRE_FLD_RUN;
-
-        rc = fld_cache_lookup(fld->lcf_cache, seq, &res);
-        if (rc == 0) {
-                *mds = res.lsr_index;
-                RETURN(0);
-        }
+       rc = fld_cache_lookup(fld->lcf_cache, seq, &res);
+       if (rc == 0) {
+               *mds = res.lsr_index;
+               RETURN(0);
+       }
 
         /* Can not find it in the cache */
         target = fld_client_get_target(fld, seq);
         LASSERT(target != NULL);
-
+       origin = target;
+again:
         CDEBUG(D_INFO, "%s: Lookup fld entry (seq: "LPX64") on "
                "target %s (idx "LPU64")\n", fld->lcf_name, seq,
                fld_target_name(target), target->ft_idx);
@@ -502,6 +494,27 @@ int fld_client_lookup(struct lu_client_fld *fld, u64 seq, u32 *mds,
                rc = fld_client_rpc(target->ft_exp, &res, FLD_QUERY, NULL);
        }
 
+       if (rc == -ESHUTDOWN) {
+               /* If fld lookup failed because the target has been shutdown,
+                * then try next target in the list, until trying all targets
+                * or fld lookup succeeds */
+               spin_lock(&fld->lcf_lock);
+
+               /* If the next entry in the list is the head of the list,
+                * move to the next entry after the head and retrieve
+                * the target. Else retreive the next target entry. */
+
+               if (target->ft_chain.next == &fld->lcf_targets)
+                       target = list_entry(target->ft_chain.next->next,
+                                           struct lu_fld_target, ft_chain);
+               else
+                       target = list_entry(target->ft_chain.next,
+                                                struct lu_fld_target,
+                                                ft_chain);
+               spin_unlock(&fld->lcf_lock);
+               if (target != origin)
+                       goto again;
+       }
        if (rc == 0) {
                *mds = res.lsr_index;
                fld_cache_insert(fld->lcf_cache, &res);
@@ -519,7 +532,7 @@ void fld_client_flush(struct lu_client_fld *fld)
 
 struct proc_dir_entry *fld_type_proc_dir;
 
-static int __init fld_mod_init(void)
+static int __init fld_init(void)
 {
        fld_type_proc_dir = lprocfs_register(LUSTRE_FLD_NAME,
                                             proc_lustre_root,
@@ -534,7 +547,7 @@ static int __init fld_mod_init(void)
        return 0;
 }
 
-static void __exit fld_mod_exit(void)
+static void __exit fld_exit(void)
 {
 #ifdef HAVE_SERVER_SUPPORT
        fld_server_mod_exit();
@@ -546,10 +559,10 @@ static void __exit fld_mod_exit(void)
        }
 }
 
-MODULE_AUTHOR("Sun Microsystems, Inc. <http://www.lustre.org/>");
-MODULE_DESCRIPTION("Lustre FLD");
+MODULE_AUTHOR("OpenSFS, Inc. <http://www.lustre.org/>");
+MODULE_DESCRIPTION("Lustre FID Location Database");
 MODULE_VERSION(LUSTRE_VERSION_STRING);
 MODULE_LICENSE("GPL");
 
-module_init(fld_mod_init);
-module_exit(fld_mod_exit);
+module_init(fld_init);
+module_exit(fld_exit);