[SOLVED] Missing full export mbdump for Saturday 20261010

MusicBrainz Development
1

The last fullexport data dump seems to have been on Wednesday:

The last SOLR datadump is from Saturday:

Could @bitmap or another developer please take a look at it?

2

Thanks, I’ve just triggered a new fullexport now. Subscriptions processing from the previous day blocked the dump again. I’m working on a patch to avoid that from happening.

4 Likes
3

Thank you for your quick response and the new fullexport:

4

@Bitmap Did the same thing happen again today (Wednesday) with the full export?

MB:

SOLR:

1 Like
5

Yes, it did. :frowning: Thanks for notifying me.

1 Like
6

Thank YOU for triggering it so quickly!

7

Sorry to bother you with this again.
Did the same thing happen again today (Saturday) with the full export, @Bitmap?

SOLR:

8

Thanks again for the quick fix:

9

@Bitmap

MB:

SOLR:

10

@Bitmap
In the meantime the MB directory Index of /pub/musicbrainz/data/fullexport/20261101-125824/ exists. However, when I run the usual fetch-dump.sh script, I receive this 404 Not Found error:

index.html                               [ <=>                                                                ]     634  --.-KB/s    in 0s      

2026-11-02 09:23:07 (292 MB/s) - ‘/media/dbdump/index.html’ saved [634]

<html>
<head><title>Index of /pub/musicbrainz/data/fullexport/</title></head>
<body>
<h1>Index of /pub/musicbrainz/data/fullexport/</h1><hr><pre><a href="../">../</a>
<a href="20261028-002228/">20261028-002228/</a>                                   28-Oct-2026 05:02       -
<a href="20261101-125824/">20261101-125824/</a>                                   01-Nov-2026 17:35       -
<a href="LATEST">LATEST</a>                                             01-Nov-2026 17:35      16
<a href="latest-is-20261101-125824">latest-is-20261101-125824</a>                          01-Nov-2026 17:35       0
</pre><hr></body>
</html>
--2026-11-02 09:23:08--  https://data.metabrainz.org/pub/musicbrainz/data/fullexport//MD5SUMS
Resolving data.metabrainz.org (data.metabrainz.org)... 138.201.203.43
Connecting to data.metabrainz.org (data.metabrainz.org)|138.201.203.43|:443... connected.
HTTP request sent, awaiting response... 404 Not Found
2026-11-02 09:23:08 ERROR 404: Not Found.

What should I do?

11

Are you still getting a 404? It’s working for me.

12

I don’t know your script, but the name of the subdirectory is missing. That should be https://data.metabrainz.org/pub/musicbrainz/data/fullexport/20261101-125824/MD5SUMS.

13

Yes, the 404 error is still there. I tried it again a few seconds ago.

I can work around the issue using replica (Fetch latest database’s replicated tables only) instead of indexed for the fetch-dump.sh script.

14

This is the official MusicBrainz-docker script to fetch the data dumps:

15

fetch-dump.sh indexed currently fails with:

ERROR 404: Not Found
https://data.metabrainz.org/pub/musicbrainz/data/fullexport//MD5SUMS

Note the double slash: the dump timestamp is empty

Cause: In indexed mode, the script reads data/solr-backups/LATEST, takes only the day part (e.g. 20261028) and looks for a directory with that exact day in data/fullexport/. Right now fullexport/ only contains 20261028-002228 and 20261101-125824, so nothing matches, DUMP_TIMESTAMP stays empty, and the script continues with a broken URL. The replica target is not affected because it uses fullexport/LATEST directly.

Recently, dumps have not always been created on the same day as the Solr backup, so a strict same-day match seems unreliable.

Possible improvements:

  1. Abort with a clear error message if DUMP_TIMESTAMP is empty, instead of requesting the non-existing .../fullexport//MD5SUMS.
  2. Fall back to fullexport/LATEST with a warning, like the replica mode does.
  3. Pick the newest database dump that is not newer than the Solr backup, instead of requiring the same day.
16

@Bitmap

MB:

SOLR: