-
Notifications
You must be signed in to change notification settings - Fork 51
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Readarr | Add info about how to run a second instance #845
Comments
I think the issue is the whole core organizational unit is the author which lives in one root library. I haven't tested trying to use the same author in two different directories but I have done it for lecturers in Sonarr for example and it will combine them if you have permission to see both. However, in the cases I do that I normally am taking my own lectures or course lectures and organizing them, They are not part of a public pre-existing group. Also they are in a different interface. Readarr seems to take 1 author and then let you choose one modality spoken or written for the whole author. You can force it book by book but it won't allow you in the UI to have both. It will treat it as an upgrade. Or at the very least it will make it very difficult to use. For now, it would be convenient until someone comes up with a better solution to allow a quick double install hosted on two separate web interfaces. Readarr:8787 ReadarrAudio:8788 is what I've selected (Speakarr) I actually still haven't got it to work. So if there progress on this at least one person would find it helpful. Lastly if there is anything about my configuration that I can post that would help debug, the situationI'd be happy to provide it. |
Yeah, at some point, either individual ebook/audiobook, author or data directory, there would need to be some Identifier in the directory name to differentiate between ebooks and audiobooks. I'm not sure whether the downloads themselves have some reliable way for Readarr to know whether it's ebook or audiobook, but otherwise e.g. a dropdown selector for the user would need to be required.
You mean the two instances? Have you tried/compared the steps I posted at the other repo? |
Yes, I changed the names around for preferences but here are the exact steps. I also tried looking at the dietpi-services but with no luck. In the end the goal is to replicate how i had it running with docker basically just two different ports 8787 and 8788 one would control epubs, the other audiobooks until there is a better solution. # Create new data directory
mkdir /mnt/dietpi_userdata/speakarr
chown -R readarr:root /mnt/dietpi_userdata/speakarr
# Setup logging to /var/log (optional)
rm -Rf /mnt/dietpi_userdata/speakarr/logs*
ln -s /var/log/speakarr /mnt/dietpi_userdata/speakarr/logs
ln -s /var/log/speakarr/logs.db /mnt/dietpi_userdata/speakarr/logs.db
ln -s /var/log/speakarr/logs.db-shm /mnt/dietpi_userdata/speakarr/logs.db-shm
ln -s /var/log/speakarr/logs.db-wal /mnt/dietpi_userdata/speakarr/logs.db-wal
# Copy service
cp /etc/systemd/system/readarr.service /etc/systemd/system/speakarr.service
# Adjust new service to use new data directory
G_CONFIG_INJECT 'Description=' 'Description=Speakarr (Readarr Audio Books)' /etc/systemd/system/speakarr.service
G_CONFIG_INJECT 'SyslogIdentifier=' 'SyslogIdentifier=speakarr' /etc/systemd/system/speakarr.service
G_CONFIG_INJECT 'LogsDirectory=' 'LogsDirectory=speakarr' /etc/systemd/system/speakarr.service
sed -i 's|/mnt/dietpi_userdata/readarr$|/mnt/dietpi_userdata/speakarr|' /etc/systemd/system/speakarr.service
# Enable and start new service
systemctl daemon-reload
systemctl enable --now speakarr |
We forgot to set the directory owner: chown readarr:root /mnt/dietpi_userdata/speakarr
systemctl restart speakarr Generally, when something is not working, check the service logs, otherwise you'll never find out the probably trivial issue: journalctl -u speakarr |
I have been trying to dig into this configuration, but it appears that tying it too closely to readarr seems to be causing a ton of issues or perhaps I'm seeing errors in the journalling that don't exist, but even though this non-dockerized version seems much smoother, this is where it all falls apart for me. So I'm not sure but it looks like there were issues with sqlite where there were so any locking errors because there are two readarr instances manipulating the same database, or something is making it seem like that. It's implied here that the database will be installed in the user directory. I wasn't sure but there were so many errors and locks in the journal so i tried making a speakarr user to further isolate the processes. https://wiki.servarr.com/en/readarr/installation#debian-ubuntu I gave speakarr access to the dietpi and readarr groups. I added this to the service to help it
I've tied a few other things, deleting the database and restarting, but it seems that even though i have a fairly large import other people have such massive ones it takes a week, yet sqlite is not locking up all the time. For now, I might have to live with throwing it into docker on another machine. Not ideal, but it seems like i can't get this working no matter what i try. In the end the imports just don't seem to show up. It would be great if this was configured properly by someone who knows the small details of whats causing so many failures. The rest is so solid. If anyone gets this working, I'd be thankful. Cheers. here is a sample of what im seeing in journal
|
Since your Speakarr has an own data directory, it also has its own database, so that cannot be the issue. I btw wouldn't use an own user but keep using the If you want to use an own user, remember to grant it write access to its data and logs directory: chown -R speakarr:root /mnt/dietpi_userdata/speakarr /var/log/speakarr |
Great, looks like the permissions of the data dir were still wrong, e.g. |
This is required if one wants to download the eboot and audiobook of the same book: MichaIng/DietPi#5004 (comment)
Alternative we can implement it as install option. But we should first check whether there is any aim to allow it with a single Readarr instance. I mean, if Readarr can download both, it cannot be so hard to allow downloading both with the same instance, giving download a related prefix/suffix or so.
The text was updated successfully, but these errors were encountered: