[prev in list] [next in list] [prev in thread] [next in thread] 

List:       ceph-devel
Subject:    [PATCH] ceph: only send metrics when the MDS rank is ready
From:       xiubli () redhat ! com
Date:       2023-05-31 11:02:01
Message-ID: 20230531110201.826061-1-xiubli () redhat ! com
[Download RAW message or body]

From: Xiubo Li <xiubli@redhat.com>

When the MDS rank is in clientreplay state these metrics requests
will be discard directly. Especially when there are a lot of known
client requests need to replay these useless metrics requests will
slow down the MDS rank.

With this we will send the metrics requests only when the MDS rank
is in active state.

URL: https://tracker.ceph.com/issues/61524
Signed-off-by: Xiubo Li <xiubli@redhat.com>
---
 fs/ceph/metric.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/fs/ceph/metric.c b/fs/ceph/metric.c
index c47347d2e84e..cce78d769f55 100644
--- a/fs/ceph/metric.c
+++ b/fs/ceph/metric.c
@@ -36,6 +36,14 @@ static bool ceph_mdsc_send_metrics(struct ceph_mds_client *mdsc,
 	s32 items = 0;
 	s32 len;
 
+	/* Do not send the metrics until the MDS rank is ready */
+	mutex_lock(&mdsc->mutex);
+	if (ceph_mdsmap_get_state(mdsc->mdsmap, s->s_mds) != CEPH_MDS_STATE_ACTIVE) {
+		mutex_unlock(&mdsc->mutex);
+		return false;
+	}
+	mutex_unlock(&mdsc->mutex);
+
 	len = sizeof(*head) + sizeof(*cap) + sizeof(*read) + sizeof(*write)
 	      + sizeof(*meta) + sizeof(*dlease) + sizeof(*files)
 	      + sizeof(*icaps) + sizeof(*inodes) + sizeof(*rsize)
-- 
2.40.1

[prev in list] [next in list] [prev in thread] [next in thread]