LU-4621 ofd: create same count of objects

when replaying the creation request (precreation) from OSP,
OFD should create the same count of objects as it created
when it processed the request initially, else it could cause
"-ENOENT" problem for the following replay requests coming
from the clients.

Change-Id: I64f61f82f6a662ccc564599468466549a55b55d0
Signed-off-by: Hongchao Zhang <hongchao.zhang@intel.com>
Test-Parameters: alwaysuploadlogs envdefinitions=SLOW=yes \
clientdistro=el6 ossdistro=el6 mdsdistro=el6 clientarch=x86_64 \
ossarch=x86_64 mdsarch=x86_64 \
clientcount=4 osscount=2 mdscount=2 \
austeroptions=-R failover=true iscsi=1 \
testlist=recovery-mds-scale
Reviewed-on: http://review.whamcloud.com/10626
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: wangdi <di.wang@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

diff --git a/lustre/ofd/ofd_dev.c b/lustre/ofd/ofd_dev.c
index 25debd1..774ec02 100644 (file)
--- a/lustre/ofd/ofd_dev.c
+++ b/lustre/ofd/ofd_dev.c
@@ -1208,6 +1208,7 @@ out_put:
 
 static int ofd_create_hdl(struct tgt_session_info *tsi)
 {
+       struct ptlrpc_request   *req = tgt_ses_req(tsi);
        struct ost_body         *repbody;
        const struct obdo       *oa = &tsi->tsi_ost_body->oa;
        struct obdo             *rep_oa;
@@ -1378,7 +1379,8 @@ static int ofd_create_hdl(struct tgt_session_info *tsi)
                               " at "LPU64"\n", ofd_name(ofd),
                               count, seq, next_id);
 
-                       if (cfs_time_after(jiffies, enough_time)) {
+                       if (!(lustre_msg_get_flags(req->rq_reqmsg) & MSG_REPLAY)
+                           && cfs_time_after(jiffies, enough_time)) {
                                CDEBUG(D_HA, "%s: Slow creates, %d/%d objects"
                                      " created at a rate of %d/s\n",
                                      ofd_name(ofd), created, diff + created,
@@ -1395,6 +1397,14 @@ static int ofd_create_hdl(struct tgt_session_info *tsi)
                                break;
                        }
                }
+
+               if (diff > 0 &&
+                   lustre_msg_get_flags(req->rq_reqmsg) & MSG_REPLAY)
+                       LCONSOLE_WARN("%s: can't create the same count of"
+                                     " objects when replaying the request"
+                                     " (diff is %d). see LU-4621\n",
+                                     ofd_name(ofd), diff);
+
                if (created > 0)
                        /* some objects got created, we can return
                         * them, even if last creation failed */

diff --git a/lustre/osp/osp_precreate.c b/lustre/osp/osp_precreate.c
index e6f81cd..7c28e1a 100644 (file)
--- a/lustre/osp/osp_precreate.c
+++ b/lustre/osp/osp_precreate.c
@@ -495,6 +495,9 @@ static int osp_precreate_send(const struct lu_env *env, struct osp_device *d)
                d->opd_pre_grow_slow = 0;
        }
 
+       body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
+       fid_to_ostid(fid, &body->oa.o_oi);
+
        d->opd_pre_last_created_fid = *fid;
        spin_unlock(&d->opd_pre_lock);