Re: [PATCH v2 3/5] ceph: check the METRIC_COLLECT feature before sending metrics

[Jeff Layton <jlayton@kernel.org>]

On Thu, 2020-06-18 at 07:59 -0400, xiubli@redhat.com wrote:
> From: Xiubo Li <xiubli@redhat.com>
> 
> Skip the MDS sessions if they don't support the metric collection,
> or the MDSs will close the socket connections directly when it get
> an unknown type message.
> 
> URL: https://tracker.ceph.com/issues/43215
> Signed-off-by: Xiubo Li <xiubli@redhat.com>
> ---
>  fs/ceph/mds_client.h | 4 +++-
>  fs/ceph/metric.c     | 8 ++++++++
>  2 files changed, 11 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h
> index bcb3892..3c65ac1 100644
> --- a/fs/ceph/mds_client.h
> +++ b/fs/ceph/mds_client.h
> @@ -28,8 +28,9 @@ enum ceph_feature_type {
>  	CEPHFS_FEATURE_LAZY_CAP_WANTED,
>  	CEPHFS_FEATURE_MULTI_RECONNECT,
>  	CEPHFS_FEATURE_DELEG_INO,
> +	CEPHFS_FEATURE_METRIC_COLLECT,
>  
> -	CEPHFS_FEATURE_MAX = CEPHFS_FEATURE_DELEG_INO,
> +	CEPHFS_FEATURE_MAX = CEPHFS_FEATURE_METRIC_COLLECT,
>  };
>  
>  /*
> @@ -43,6 +44,7 @@ enum ceph_feature_type {
>  	CEPHFS_FEATURE_LAZY_CAP_WANTED,		\
>  	CEPHFS_FEATURE_MULTI_RECONNECT,		\
>  	CEPHFS_FEATURE_DELEG_INO,		\
> +	CEPHFS_FEATURE_METRIC_COLLECT,		\
>  						\
>  	CEPHFS_FEATURE_MAX,			\
>  }
> diff --git a/fs/ceph/metric.c b/fs/ceph/metric.c
> index 27cb541..4267b46 100644
> --- a/fs/ceph/metric.c
> +++ b/fs/ceph/metric.c
> @@ -127,6 +127,14 @@ static void metric_delayed_work(struct work_struct *work)
>  			continue;
>  		}
>  
> +		/*
> +		 * Skip it if MDS doesn't support the metric collection,
> +		 * or the MDS will close the session's socket connection
> +		 * directly when it get this message.
> +		 */
> +		if (!test_bit(CEPHFS_FEATURE_METRIC_COLLECT, &s->s_features))
> +			continue;
> +
>  		/* Only send the metric once in any available session */
>  		ret = ceph_mdsc_send_metrics(mdsc, s, nr_caps);
>  		ceph_put_mds_session(s);

This should probably be moved ahead of, or folded into the previous
patch to prevent a regression should someone land in between them when
bisecting.
-- 
Jeff Layton <jlayton@kernel.org>