Issue/Introduction

This article covers two issues during ZStone 4.2.8 initialization: the host addition workflow fails at 67% progress, and the system disk capacity is displayed as 0.

Symptoms

• The host addition workflow fails at 67% progress.
• The system disk capacity is displayed as 0.

Environment

• Product line: ZCF
• Component: ZStone
• Version: 4.2.8

Possible causes

The two issues have the following causes:

1. The public network and the cluster network use different IP addresses in the same subnet. When the MON is created, the node name resolves to the public network IP, which causes MON creation to fail and MonCephConfig to time out.
2. The system disk uses the LVM partition format, which is not supported by this ZStone version. As a result, the system disk capacity is displayed as 0.

Resolution

1. Check the MonCephConfig task

• Check: Inspect the Temporal log and confirm whether a MonCephConfig timeout occurred.
• Command:

  ceph -s

• Abnormal result: The command hangs, which indicates that the MON was not created successfully.

2. Check the MON network configuration

• Action: Stop the MON service, extract the monmap, and inspect it.
• Commands:

  ceph-mon -i `hostname` --extract-monmap /tmp/monmap
  monmaptool /tmp/monmap --print

• Check: Confirm whether the cluster network IP in the monmap matches the expected configuration.
• If abnormal: If the public network and the cluster network use the same subnet, replan the networks and initialize the cluster with IP addresses from different subnets.

3. Check the system disk partition format

• Command:

  ls /dev/disk/by-id | grep -E 'wwn|virtio|nvme-eui' | grep -v 'part'

• Check: Confirm whether the system disk uses the LVM partition format.
• If abnormal: Reinstall the operating system without using the LVM partition format.