From: Bruno Faccini <bruno.faccini@intel.com>
Date: Fri, 3 Feb 2017 16:55:37 +0000 (+0100)
Subject: LU-9075 mdt: avoid race causing mdt_coordinator_cb() err msgs
X-Git-Tag: 2.9.57~79
X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=commitdiff_plain;h=d1535dc90b01770e56a0c79c7bb1e7c9cd8f1c6a

LU-9075 mdt: avoid race causing mdt_coordinator_cb() err msgs

This patch mainly moves mdt_agent_record_update() call before
mdt_cdt_remove_request() in mdt_hsm_update_request_state(), to
avoid the frequent couple of
"(mdt_coordinator.c:1473:mdt_hsm_update_request_state()) ...
Cannot find running request for cookie ..."
and
"(mdt_coordinator.c:339:mdt_coordinator_cb()) ...
cannot cleanup timed out request ..."
error msgs, likely to concern active requests that have completed
and thus that have already been removed from memory in
mdt_hsm_update_request_state() (using mdt_cdt_remove_request() and
in the context of a MDT thread handling CT's MDS_HSM_PROGRESS
requests), but the corresponding action LLOG record update is stuck
awaiting for CDT to give-back cdt_llog_lock in
mdt_agent_record_update().

Others related but minor changes are, use of arr_req_change instead of
arr_req_create to more accuratelly determine if a request exceeds the
timeout, and change main debug msg in mdt_hsm_update_request_state()
to reflect if action LLOG record update will occur or not.

Signed-off-by: Bruno Faccini <bruno.faccini@intel.com>
Change-Id: I043813f1ff11a7e9e99c534fa8560a35e2c52543
Reviewed-on: https://review.whamcloud.com/25243
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Henri Doreau <henri.doreau@cea.fr>
Reviewed-by: Quentin Bouget <quentin.bouget@cea.fr>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
---

diff --git a/lustre/mdt/mdt_coordinator.c b/lustre/mdt/mdt_coordinator.c
index c563284..445bf7d 100644
--- a/lustre/mdt/mdt_coordinator.c
+++ b/lustre/mdt/mdt_coordinator.c
@@ -284,7 +284,7 @@ static int mdt_coordinator_cb(const struct lu_env *env,
 		 */
 		car = mdt_cdt_find_request(cdt, larr->arr_hai.hai_cookie, NULL);
 		if (car == NULL) {
-			last = larr->arr_req_create;
+			last = larr->arr_req_change;
 		} else {
 			last = car->car_req_update;
 			mdt_cdt_put_request(car);
@@ -1427,15 +1427,14 @@ int mdt_hsm_update_request_state(struct mdt_thread_info *mti,
 
 		rc = hsm_cdt_request_completed(mti, pgs, car, &status);
 
-		/* remove request from memory list */
-		mdt_cdt_remove_request(cdt, pgs->hpk_cookie);
-
-		CDEBUG(D_HSM, "Updating record: fid="DFID" cookie=%#llx"
-			      " action=%s status=%s\n", PFID(&pgs->hpk_fid),
-		       pgs->hpk_cookie,
+		CDEBUG(D_HSM, "%s record: fid="DFID" cookie=%#llx action=%s "
+			      "status=%s\n",
+		       update_record ? "Updating" : "Not updating",
+		       PFID(&pgs->hpk_fid), pgs->hpk_cookie,
 		       hsm_copytool_action2name(car->car_hai->hai_action),
 		       agent_req_status2name(status));
 
+		/* update record first (LU-9075) */
 		if (update_record) {
 			int rc1;
 
@@ -1451,6 +1450,10 @@ int mdt_hsm_update_request_state(struct mdt_thread_info *mti,
 				       pgs->hpk_cookie);
 			rc = (rc != 0 ? rc : rc1);
 		}
+
+		/* then remove request from memory list (LU-9075) */
+		mdt_cdt_remove_request(cdt, pgs->hpk_cookie);
+
 		/* ct has completed a request, so a slot is available, wakeup
 		 * cdt to find new work */
 		mdt_hsm_cdt_wakeup(mdt);