Closed whispy closed 6 years ago
Thanks for the report. I'm getting feedback left and right, so it might take a bit until I can take a look.
I'm running NZBHydra 1, version 0.2.231 (see: https://vgy.me/pEfIX2.png).
Yeah sorry, I didn't see that in the headline ;-)
There's a column missing in the database that was added with 0.2.151. I can't really explain why it's not there. That should not be possible. That could mean that there are any number of possible database states out there. Oh my god.
I'll have to release a new update for v1 which makes sure everything is expected but that will take time. Considering how many reports I'm getting for v2 you're better off waiting a bit ;-)
Okay, thanks. Is that also why the settings.cfg
import fails? I can do without my old database, but manually recreating all the settings and indexers in the UI would be a bit of a pain (but doable)
EDIT: I can send you my DB file too, if you'd like.
Please post your debug infos. I didn't really expect so many (or any issues) with the migration.
The settings should be migrated even if the database migration fails.
Please post your debug infos from v2 (http://127.0.0.1:5076/system/bugreport)
I'm seeing
Error while reading old settings file
java.io.IOException: Is a directory
Make sure that the settings file you provided is correct.
That was from a previous attempt where I just entered /data/nzbhydra-old-migration
into the UI to see if it needed the filename or not. Other attempts in the debuginfos should have the full file paths.
EDIT: And I was doing the manual migration, as my V1 was running in a Docker container.
Hm shit, the migration will fail when the database migration doesn't work.
As a workaround you could shut down your v1 instance, move the database files and start v1 again. Then try to migrate again. The v1 database should be reinitialized and match the expectations of the migration process.
I really want to know what your database looks like and why it looks like that. The missing columns were added 2016 and Hydra shouldn't even work without them...
I figured it out, and it was me being silly — apologies for wasting your time.
I noticed when moving the database files, as you wrote earlier, that I was mistakenly using an older DB file from the end of 2016 (probably left over from some previous troubleshooting) that was in the top-level of my /hydra
folder, and not the most up-to-date database. I corrected this and used the most recent DB file, and the migration worked fine.
Sorry for any potential stress! Keep up the awesome work!
Sent you a small donation to apologize :P
Aaargh :-D
Well, I'm really glad that you figured it out quickly. I'll add code that checks the database version in the database file directly.
Thanks for the donation.
Steps:
settings.cfg
andnzbhydra.db
into a sub-folder of NZBHydra2's/data
volume.System
>Control
>Migrate from NZBHydra 1
.NZBHydra 1 is running on another machine or in a container
.Start Migration
.Results:
Config
>Indexers
. Other settings also seem to have failed to import.java.sql.SQLException: no such column: 'author'
(more info in log section below).Note: I have not yet tried running with
--xmx 512M
, as Synology's Docker UI makes this difficult and my log files make it seem like this is not the issue. However, if you request, I will figure out how to run it with this argument.Log (debug output enabled and indexer names redacted):
Please let me know if I can provide any more info or testing. Thanks for putting all this effort into the rewrite!