[...] index files outside of expected directory scheme

[Fantomas]

Hello community,

when performing garbage collection on a datastore, I have observed following log message:

2022-11-08T11:38:08+01:00: marked 100% (414 of 414 index files)
2022-11-08T11:38:08+01:00: found (and marked) 414 index files outside of expected directory scheme
2022-11-08T11:38:08+01:00: Start GC phase2 (sweep unused chunks)
2022-11-08T11:38:35+01:00: processed 1% (7052 chunks)

I assume a relation to manual introduction of namespaces after initial backups, my datastore's directory layout looks as follows:

# tree -a -d -I ???? -L 4
.
├── .chunks
└── ns
    ├── namespace-a
    │   ├── ct
    │   │   └── 203
    │   └── vm
    │       └── 201
    └── namespace-b
        ├── ct
        │   ├── 102
        │   ├── 111
        │   └── ...
        └── vm
            ├── 101
            ├── 103
            └── ...

My assumption was that a central .chunks folder above specific namespaces is the desired behavior. Could anybody confirm / comment on that?

Thanks and best regards,
Jan

 

[Neobin]

Checked on my PBSs and can confirm:
Since the switch (via the GUI using local remote sync; so no manual intervention on the filesystem-side from me) to namespaces, I also have those: "found (and marked) <XYZ> index files outside of expected directory scheme" entries in every GC-log.
My folder structure is the same as yours.

 

[Neobin]

Sorry to bump this thread, but I would be happy, if someone from the Staff could confirm, that this message does not come from a misconfiguration and is nothing to "worry" about and if this is the case, if it might make sense to rephrase it?

 

[fabian]

yes, this seems to have been overlooked and should be benign (provided there are no other directories / nested datastores that don't conform to the "namespace" scheme)

 

[fabian]

filed https://bugzilla.proxmox.com/show_bug.cgi?id=4357 for tracking

 

[Neobin]

Thank you very much.

 

[ITT]

Oh, same here...