Skip to content

Unraid webGUI /Main returns 500 after adding disk to global share until reboot #2570

@kubbur

Description

@kubbur

Summary

After adding a disk to a global share, the Unraid webGUI became partially broken until a full reboot.

Symptoms included:
• login page still worked
• /Main returned 500 Internal Server Error
• emhttpd would not stay up reliably
• /usr/local/emhttp/state was missing
• newer Unraid API/UI components complained about missing emhttp state

A normal reboot restored the system.

Environment
• Unraid version: 7.2.0
• Access method: local webGUI + SSH
• Reverse proxy/Caddy also present, but issue reproduced directly on internal ip:port
• Docker services otherwise largely continued working

Probable trigger

A disk was added to a global share, and the system was not rebooted afterward.

This is the most likely trigger, but not conclusively proven.

Observed behavior

User-facing symptoms
• Unraid login page loaded normally
• after login, /Main returned HTTP 500
• direct access to internal ip:port allowed login, then failed on main UI
• other services on the host remained reachable

Runtime observations

emhttpd was initially running, but /usr/local/emhttp/state was missing entirely.

Relevant state check:

ls -l /usr/local/emhttp/state/var.ini
ls -l /usr/local/emhttp/state/disks.ini
ls -l /usr/local/emhttp/state/shares.ini

All expected state files were missing because /usr/local/emhttp/state itself did not exist.

Relevant logs

Unraid API / graphql-api.log repeatedly showed missing state:

Failed loading state file from "/usr/local/emhttp/state/var.ini" with "Invalid File Path: /usr/local/emhttp/state/var.ini, or Extension: ini"
Failed loading state file from "/usr/local/emhttp/state/devs.ini" with "Invalid File Path: /usr/local/emhttp/state/devs.ini, or Extension: ini"
Failed loading state file from "/usr/local/emhttp/state/network.ini" with "Invalid File Path: /usr/local/emhttp/state/network.ini, or Extension: ini"
Failed loading state file from "/usr/local/emhttp/state/nginx.ini" with "Invalid File Path: /usr/local/emhttp/state/nginx.ini, or Extension: ini"
Failed loading state file from "/usr/local/emhttp/state/shares.ini" with "Invalid File Path: /usr/local/emhttp/state/shares.ini, or Extension: ini"
Failed loading state file from "/usr/local/emhttp/state/disks.ini" with "Invalid File Path: /usr/local/emhttp/state/disks.ini, or Extension: ini"
Failed loading state file from "/usr/local/emhttp/state/users.ini" with "Invalid File Path: /usr/local/emhttp/state/users.ini, or Extension: ini"
Failed loading state file from "/usr/local/emhttp/state/sec.ini" with "Invalid File Path: /usr/local/emhttp/state/sec.ini, or Extension: ini"
Failed loading state file from "/usr/local/emhttp/state/sec_nfs.ini" with "Invalid File Path: /usr/local/emhttp/state/sec_nfs.ini, or Extension: ini"

Also present:

Missing config values: store.emhttp.var.version is not set, store.emhttp.var.flashGuid is not set
Failed to compare Unraid version - missing unraid version

Manual emhttpd stop/start via /usr/local/sbin/emhttp showed:

emhttpd: Unraid(tm) System Management Utility version 7.2.0
emhttpd: unclean shutdown detected
emhttpd: get_diskFsSizes 8588 diskFsType.0 not found

After that, emhttpd exited and /usr/local/emhttp/state was still not recreated.

Troubleshooting performed

Tried before reboot:
• verified issue directly on Unraid webGUI
• disabled/removed live Unassigned Devices plugin directories from /usr/local/emhttp/plugins
• removed ca.mover.tuning from live plugin tree
• restarted nginx / php-fpm / unraid-api multiple times
• manually stopped/started web stack using /usr/local/sbin/emhttp
• checked syslog, nginx logs, php-fpm logs, and graphql-api logs
• verified emhttpd socket existed
• verified emhttpd process path: /usr/libexec/unraid/emhttpd

None of the above restored /usr/local/emhttp/state or fixed /Main.

Resolution

A normal full reboot fixed the problem.

After reboot:
• webGUI recovered
• /Main worked again
• issue considered resolved

Expected behavior

If adding a disk to a global share requires a reboot or service restart to safely rebuild management state, Unraid should either:
• do that automatically, or
• clearly warn that a restart/reboot is required

It should not leave the system in a state where:
• /Main returns 500
• /usr/local/emhttp/state disappears
• emhttpd fails to remain available

Impact
• webGUI partially unusable
• troubleshooting had to be done over SSH
• issue looked initially like plugin/API corruption but was cleared by reboot

Workaround

Reboot the server after adding the disk to the global share.

Notes

This report does not prove the exact internal root cause, but the timeline strongly suggests the share/disk change caused management state inconsistency until reboot.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions