Node expecting cache dir that doesnt exist 2.27.02

[orcutt989]

Windows Server, Linux Node (Kubernetes)
Tdarr Version 2.27.02

Jc6YDXxPMD-log.txt

When I unpause the node the logs fill up with errors saying it can't access a directory that doesnt exist when processing a transcode. Its as if its looking for a cache that never existed.

│ [2025-01-06T08:28:53.925] [ERROR] Tdarr_Node - Error: ENOENT: no such file or directory, access '/mnt/smb/d/tdarr_cache/tdarr-wo │
│ rkDir2-4Be3WQaiA'                                                                                                                │
│     at Object.accessSync (node:fs:254:3)                                                                                         │
│     at c (/app/Tdarr_Node/srcug/workers/worker1.js:1:27644)                                                                      │
│     at preProcessFile (/app/Tdarr_Node/srcug/workers/worker1.js:1:29462){                                                        │
│   "errno": -2,                                                                                                                   │
│   "syscall": "access",                                                                                                           │
│   "code": "ENOENT",                                                                                                              │
│   "path": "/mnt/smb/d/tdarr_cache/tdarr-workDir2-4Be3WQaiA"                                                                      │
│ }                                                                                                                                │
│ [2025-01-06T08:28:55.359] [ERROR] Tdarr_Node - Error: ENOENT: no such file or directory, access '/mnt/smb/d/tdarr_cache/tdarr-wo │
│ rkDir2-ioOQ5YZnOT'                                                                                                               │
│     at Object.accessSync (node:fs:254:3)                                                                                         │
│     at c (/app/Tdarr_Node/srcug/workers/worker1.js:1:27644)                                                                      │
│     at preProcessFile (/app/Tdarr_Node/srcug/workers/worker1.js:1:29462){                                                        │
│   "errno": -2,                                                                                                                   │
│   "syscall": "access",                                                                                                           │
│   "code": "ENOENT",                                                                                                              │
│   "path": "/mnt/smb/d/tdarr_cache/tdarr-workDir2-ioOQ5YZnOT"                                                                     │
│ }    

The cache dir is completely empty. Not sure why its looking for folders that dont exist? My mounts are accessible from the node.

root@tdarr-node-f984bd885-sgntx:/# ls /mnt/smb/d/tdarr_cache/
root@tdarr-node-f984bd885-sgntx:/#

Here are my pathTranslators for this path.

        {
          "server": "D:/tdarr_cache",
          "node": "/mnt/smb/d/tdarr_cache/"
        },

Here is the cache path in the Library

I cleared all the library DBs, restarted the server, rejoined nodes and still sam error when it goes to process a transcode.

I verified access to the filesytem by shelling into the tdarr node pod and creating a file.

root@tdarr-node-f984bd885-qspnk:/mnt/smb/d/tdarr_cache# touch test
root@tdarr-node-f984bd885-qspnk:/mnt/smb/d/tdarr_cache# ls
test

[HaveAGitGat]

This is usually a permissions or path issue. Tdarr server automatically creates and cleans up the sub cache directories (like tdarr-workDir2-4Be3WQaiA). In this case the node can't access that directory.

Create this path on the server: D:/tdarr_cache/test, then check it exists by going to /mnt/smb/d/tdarr_cache/test on the node.

[orcutt989]

Done

root@tdarr-node-f984bd885-q8d5p:/# ls -lah /mnt/smb/d/tdarr_cache/
total 12K
drwxr-xr-x 2 root root   0 Jan 15 20:28 .
drwxr-xr-x 2 root root 12K Jan 15 20:27 ..
-rwxr-xr-x 1 root root   0 Jan 15 20:28 test

I can even write to it from the node

echo "test" > /mnt/smb/d/tdarr_cache/test
cat /mnt/smb/d/tdarr_cache/test
test

And I can remove it from the node as well

root@tdarr-node-f984bd885-q8d5p:/# ls /mnt/smb/d/tdarr_cache/
test
root@tdarr-node-f984bd885-q8d5p:/# rm /mnt/smb/d/tdarr_cache/test
root@tdarr-node-f984bd885-q8d5p:/# ls /mnt/smb/d/tdarr_cache/
root@tdarr-node-f984bd885-q8d5p:/#

[HaveAGitGat]

I assume the process permissions when it creates the folder are slightly different to when you do it manually and the node can't access it in that case.

Have you tried running the stack in a flow?

You can use a Wait plugin (set it to a e.g 300 seconds) so you can then inspect the cache folder and try and create a file: