FailedProxy/Varaverk
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Point the WebGUI symlink at the actual install, and report this host's identity rather than HOST1's

Browse Source 
The .plg hardcoded the flash path and runs every boot, so an appdata node served a stale copy that never receives pulls — fixes appeared to do nothing, indefinitely.
This commit is contained in:
Gmer4Lfe
2026-08-17 11:52:33 -04:00
parent 620d7eb6f1
commit 7ae510ed28
3 changed files with 47 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Plugin/varaverk.plg
+16 -1
View File
@@ -52,7 +52,22 @@
<![CDATA[
#!/bin/bash
WEB_DIR="/usr/local/emhttp/plugins/varaverk"
SRC="/boot/config/plugins/varaverk/Plugin/unraid"
# SRC follows the install, it is not assumed. This was hardcoded to the flash path while this
# block runs on EVERY boot, so an appdata-mode node had its WebGUI pointed back at the flash copy
# every time it started — a copy that stops receiving pulls the moment SCRIPTS_DIR moves. The
# result is a host where git pull reports success, the files on disk are current, and the pages
# served are whatever the flash copy last had. Fixes appear to do nothing, indefinitely.
#
# varaverk.cfg stays on flash regardless of layout, which is what makes it readable this early.
SRC=""
if [[ -f /boot/config/plugins/varaverk/varaverk.cfg ]]; then
_sd=$( . /boot/config/plugins/varaverk/varaverk.cfg >/dev/null 2>&1 || true
printf '%s' "${SCRIPTS_DIR:-}" )
[[ -n "$_sd" && -d "$_sd/Plugin/unraid" ]] && SRC="$_sd/Plugin/unraid"
fi
[[ -z "$SRC" ]] && SRC="/boot/config/plugins/varaverk/Plugin/unraid"
[[ -L "$WEB_DIR" ]] && rm -f "$WEB_DIR"
[[ -d "$WEB_DIR" ]] && rm -rf "$WEB_DIR"
ln -sf "$SRC" "$WEB_DIR"