RHEL7: NFSv4 client hangs with tcpdump showing repeated sequence of OPEN / TEST_STATEID all completing with NFS_OK but SEQUENCE reply contains SEQ4_STATUS_RECALLABLE_STATE_REVOKED bit set
Issue
- NFS4 client hangs and must be rebooted to recover.
- Seeing many
kernel: NFS: nfs4_reclaim_open_state: Lock reclaim failed!
messages as described in https://access.redhat.com/solutions/1117763 - Taking a tcpdump, it is seen that the NFS4 state manager sends TEST_STATEID and receiving SEQ4_STATUS_RECALLABLE_STATE_REVOKED in a tight loop, > 5000 times per second.
2279761 07:28:28.725640 10.1.1.1 → 10.1.1.2 NFS V4 Call TEST_STATEID
2279762 07:28:28.725674 10.1.1.2 → 10.1.1.1 NFS 0x00000040 V4 Reply (Call In 2279761) TEST_STATEID
status flags: 0x00000040, SEQ4_STATUS_RECALLABLE_STATE_REVOKED
- In /var/log/messages many 'suppressed' messages like
Aug 1 13:26:22 nfs-client kernel: nfs4_reclaim_open_state: 2265 callbacks suppressed
Environment
- Red Hat Enterprise Linux 7 (NFS client and server)
- seen on 3.10.0-862.*.el7 and 3.10.0-957.*el7 kernels
- NFS4.1 or above
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.