ceph warn oldest client/flush tid

    health: HEALTH_WARN
            1 clients failing to advance oldest client/flush tid


[root@ceph0001 ~]# ceph health detail
HEALTH_WARN 1 clients failing to advance oldest client/flush tid
MDS_CLIENT_OLDEST_TID 1 clients failing to advance oldest client/flush tid
    mdsceph0002(mds.0): Client node0001 failing to advance its oldest client/flush tid client_id: 2169313

　　

1.找到对应的id

 

[root@ceph0002 ~]# ceph daemon mds.ceph0002 session  ls|grep 2169313 -A 10
        "id": 2169313,
        "num_leases": 0,
        "num_caps": 39,
        "state": "open",
        "request_load_avg": 150519,
        "uptime": 16855.113280135,
        "replay_requests": 0,
        "completed_requests": 2287874,
        "reconnecting": false,
        "inst": "client.2169313 1.1.1.1:0/3642728200",
        "client_metadata": {
            "features": "0000000000001fff",
            "ceph_sha1": "xxxxxxxxxxxxxxxxxxxxxxxx",
            "ceph_version": "ceph version 14.2.3 (xxxxxxxxxxxxxxxxxxxxx) nautilus (stable)",
            "entity_id": "kubernetes-dynamic-user-baff381e-7a97xxxxxxxxxxxxxxxa",
            "hostname": "node0001",
            "mount_point": "/var/lib/kubelet/pods/c6e01082-f513-493d-8b99-2cc8ef8f6df7/volumes/kubernetes.io~cephfs/pvc-6de4a2c4-29cc-4335-93ff-9fae1e321201",
            "pid": "26838",
            "root": "/1/2/3/kubernetes-dynamic-pvc-baff36eb-7a97-11ec-a5ce-da4d9b86e22a"
        }

　　

2.找到对应的pod

kubectl  get pod -A -o yaml |grep c6e01082-f513-493d-8b99-2cc8ef8f6df7 -C 10

　　

3.删除pod，客户端主动释放

 

4删不掉的情况下可以ceph 释放该链接

 ceph tell mds.0 client evict id=2169313