A Network File System (NFS) client is part of a kernel. After you install an NFS client, errors may occur due to pre-existing bugs in the kernel. To ensure the stability of NFS clients, we recommend that you use the kernel versions that are verified by Alibaba Cloud.

Recommended Linux versions

To ensure system stability, we recommend that you use Alibaba Cloud official images whose kernel versions are strictly tested and verified. For more information, see Recommended kernel images.

NFS response failures due to a kernel bug in network stacks (priority: high)

An NFS response failure that is caused by a network stack error occurs if an NFS server fails to process a request that is repeatedly sent from a kernel. This issue may occur in the following kernel versions: 2.6.32-296 to 2.6.32-696.10.1, except for 2.6.32-696.10.1.

If the "operation failed" message appears, we recommend that you restart the Elastic Cloud Service (ECS) instance on which the client resides. For more information, see RHEL6.9:NFSv4 TCP transport stuck in FIN_WAIT_2 forever.

NFS response failures due to a kernel bug (priority: high)

  • An NFS response failure occurs if an NFS server failover occurs after you start a connected NFS client or read data from or write data to a file system by using the NFS client. The failover results in a deadlock. This issue may occur in the following kernel versions:
    • RHEL 6 or CentOS 6 2.6.32-696.3.1.el6
    • RHEL 7 or CentOS 7 3.10.0-229.11.1.el7 and earlier versions
    • Ubuntu 15.10: Linux 4.2.0-18-generic

    If the "operation failed" message appears, we recommend that you restart the ECS instance on which the client resides. For more information, see RHEL7:NFSv4 client loops with WRITE/NFS4ERR_STALE_STATEID - if NFS server restarts multiple times within the grace period.

  • An NFS response failure occurs if network partition or network jitter occurs or if the NFS client fails to provide appropriate solutions to the returned error codes. Network partition or network jitter results in repeated network connections. If this issue occurs, the "bad sequence-id" error message repeatedly appears in system logs. This issue may occur in the following kernel versions:
    • RHEL 6 or CentOS 6 2.6.32-696.16.1.el6 and earlier
    • RHEL 7 or CentOS 7 3.10.0-693.el7 and earlier

    If the "operation failed" message appears, we recommend that you restart the ECS instance on which the client resides. For more information, see RHEL6/RHEL7:NFS4 client receiving NFS4ERR_BAD_SEQID drops nfs4 stateowner resulting in infinite loop of READ/WRITE+NFS4ERR_BAD_STATEID.

  • An NFS response delay or failure occurs when you run the ls command, commands with the * or ? wildcard, or other commands that require directory traversal. This issue occurs in the CentOS and RedHat 5.11.x kernel versions.

    To fix the issue, we recommend that you upgrade the kernel to the latest version.

Unsupported chown command and system API calls (priority: low)

If the kernel version is 2.6.32, the chown command and system API calls are not supported on an NFS client.

ls command in an infinite loop (priority: low)

  • If the kernel version is 2.6.32-696.1.1.el6 or earlier, the ls command runs in an infinite loop. This issue occurs if you create or delete files or subdirectories when the ls command is running.

    To fix the issue, we recommend that you upgrade the kernel to the latest version.

  • If the kernel version of your Linux is 4.18.0-305.12.1, commands (such as ls) that require directory traversal run in an infinite loop. To fix this issue, we recommend that you upgrade the kernel to 4.18.0-305.19.1. For more information, see kernel-4.18.0-305.19.1.el8_4.x86_64.