vsanfs NFS returns NFS4ERR_RESOURCE for LOCK call after network reconnection

Products

• VMware vSAN 8.x

Issue/Introduction

After network disconnection, NFS client tries to reconnect to vsanfs NFS, but it returns NFS4ERR_RESOURCE for Lock call and eventually IO fails for a moment.

You might see below entry from ganesha.log.

YYYY-MM-DDThh:mm:ssZ : epoch ######## : ######.######.###### : ganesha.nfsd-104[::ffff:###.###.###.###] [svc_3181] 515 :nfs4_op_lock :NFS4 LOCK :LOCK failed to create new lock owner Lock: obj=0x7f0274002c78, fileid=487297011706226969, type=READ , start=0x0, end=0xffffffffffffffff, owner={STATE_OPEN_OWNER_NFSV4 0x7f02b0003930: clientid={0x7f02841031f0 ClientID={Epoch=######## Counter=0x00000881} CONFIRMED Client={0x7f0284001ba0 name=(22:Linux NFSv4.1 ######) refcount=1} t_delta=0 reservations=2 refcount=10} owner=(24:0x6f70656e2069643a000000310000000000000524cb97b8e5) confirmed=1 seqid=0 refcount=2}

Environment

VMware vSAN 8.X

Cause

This is a known issue for vsanfs NFS. Broadcom engineering team is aware of this issue and the fix will be provided in the future release.

Resolution

There is currently no workaround for this issue. However, the Broadcom engineering team is aware of it and will provide a fix in a future release.

Because this only occurs following an NFS session reconnection, it does not impact normal operating conditions.

Additional Information

日本語版: vsanfs NFS がネットワークの再接続後に LOCK call に対し NFS4ERR_RESOURCE のエラーを返す