Question: Calibre upgrade: Large version jump

[haertig]

I have been running Calibre on my Linux system. I just now realized that I am running a very old copy. I'm running a recent version of the Linux distro, but I guess the repos for applications is a bit outdated, at least for Calibre.

I am currently running Calibre version 2.55

The newest version available from the Calibre website is 3.9.0

That's quite a jump. I known my way around Linux, and would have no problem building from scratch, manually fixing dependencies, etc. So the method in which I install the newer version doesn't matter. Using a package, or compiling - doesn't matter to me.

However, my concern is that i'd like to keep all my existing metadata. I've done a lot of work setting tags, downloading book covers that I like, etc. I don't want to lose that. I assume I would lose it if I uninstall the existing 2.55 Calibre before installing the newer 3.9.0 version. Possibly not, but possibly so, therefore I don't want to chance it without researching the best upgrade path first.

Can I just install the newest Calibre over the top of my existing version 2.55? Usually this is OK with many applications, but some times big jumps in versions (like I need to do) can cause problems. Of course, my entire Calibre directory structure is backed up. Is that sufficient if a bare metal recovery ends up being required following some disaster? Or does Calibre place information in other places (besides the main Calibre directory structure) that I need to first identify and make sure that is backed up? I know the books themselves are safely backed up, but maybe not the metadata, depending on how/where Calibre stores that.

Thanks in advance!

[theducks]

Calibre will do fine with the upgrade.

Do make a backup of your Library folder, just in case...
The Database structure has been pretty stable for years. It is the features that have soared.

Very few Linux distros are anywhere near current.

[BetterRed]

Originally Posted by theducks

Calibre will do fine with the upgrade.

The change from version 2 to 3 is primarily related to the server, the library manager didn't change, apart from some cosmetic changes in the tag browser, status and search bars.

If you use the server, I suggest you read the relevant section in the manual and recent release notes. Here's a start calibre - New in calibre 3.0

BR

[haertig]

Thanks for the quick replies. I read the webpage on what's new in version 3.0. Unexpectedly, that actually gave me pause regarding doing the upgrade.

I was just beginning to investigate use of the content server. I played with this for a while in my version 2.55. It looked really neat, but some of the documentation I found did not match what I was seeing on-screen. That's when I figured out I was running an old version. Specifically, I was looking at running the content server full-time on my Linux box. But lo and behold, it appears that something new in version 3 says you can't run a server full time, and use the local Calibre interface at the same time. But you can run the server from within the user interface. That means I'd have to leave all of Calibre running full time. Or kill the server, run the user interface, then restart the server. Neither of those approaches sounds appealing. It just doesn't seem "clean".

Hmmm, that kind of defeats the purpose of what I wanted to do (a full time running server). So it appears the trade-off is getting a server with a few more features, but giving up the ability to run that server all the time as a (hopefully smaller memory footprint) daemon process. I'll have to research more, but I think I might actually prefer the older version that (apparently) allows both the content server and the main Calibre interface to run simultaneously, independent of each other. I don't need, and actually don't WANT, for others to attach to the content server and be able to modify anything. I want non-destructive read-only on the content server. This appears to be what version 2 supported, but maybe not version 3 (although I suppose there may be configuration options to make the server read-only).

Again, I'll have to do more research before I decide. But if the main difference between version 2 and 3 is the server, and the new server is less of what I want than the old, it may be best to just stick with the older version I'm currently running.

[haertig]

What I may do, and probably WILL do after I think more about it, is to run TWO Calibres.

I would run my main Caliber on my desktop computer. This is what I am doing now. Then I would set up a cron job to rsync the Calibre directory over to my Raspberry Pi3, where I have 4tb of available storage. Then run a SECOND Calibre, as a content server, off of the Pi - serving up it's rsynced copy of my main Calibre database. My Calibre library is already preset over on that Pi - as a set of backups via UrBackup. But I wouldn't want to be running a content server against my backup - that could modify/corrupt the backup which defeats it's purpose.

I kind of like this idea the more I think about it. I am already using that Pi to run my personal NextCloud server, and a UrBackup server. Might as well add a Calibre server to that mix. I thought about adding my Plex Media Server to the Pi also, but decided against that as Plex can be more resource intensive, especially if it needs to transcode something. Anyway, NextCloud + UrBackup + Calibre sounds like a handy little Raspberry Pi set up.

I'd just have to do some testing to make sure a running Calibre content server wouldn't get confused by changes being made to it's data directory by an asynchronous rsync. May have to shut down the content server during the rsyncs. But could possibly still have problems. If I snapshot my main Calibre database prior to the rsync, to ensure data integrity, I think it should work. In theory.

[mbovenka]

Originally Posted by haertig

I would run my main Caliber on my desktop computer. This is what I am doing now. Then I would set up a cron job to rsync the Calibre directory over to my Raspberry Pi3, where I have 4tb of available storage. Then run a SECOND Calibre, as a content server, off of the Pi - serving up it's rsynced copy of my main Calibre database.

That's pretty much what I do using COPS on my NAS (a Synology DS213J, so no full Calibre) and FreeFileSync.

[JSWolf]

Originally Posted by theducks

Calibre will do fine with the upgrade.

Do make a backup of your Library folder, just in case...
The Database structure has been pretty stable for years. It is the features that have soared.

Very few Linux distros are anywhere near current.

There's no need to backup the Library just in case.

[kacir]

Originally Posted by haertig

I have been running Calibre on my Linux system. I just now realized that I am running a very old copy. I'm running a recent version of the Linux distro, but I guess the repos for applications is a bit outdated, at least for

The configuration data is stored at
~/.config/calibre/
Make a backup. This contains all plugins and configutations.
Configuration of custom columns and other library-related things are stored at the Calibre Library directory. I want to believe you already have a good backup of this directory. Make another one ;-)

The thing is, that the old Calibre is most probably installed somewhere in /usr/local/bin or in other place where other GUI programs installed by your distro are located.
When you install Calibre Library from the Calibre.org site, it will be installed to the /opt/calibre directory. So you do not have to uninstall anything, the old Calibre will remain in place.

Before installing the new version of Calibre go to terminal and type
which calibre
to see where your Calibre as packaged by a distribution is installed. You will most probably see something like
/usr/bin/calibre
Now, have a look at this file. It is most probably just a symlink pointing you to the "real" calibre binary
ls -l /usr/bin/calibre

You should always install Calibre from the official site
https://calibre-ebook.com/download_linux
The vast majority of Linux distributions do not keep up with the fast pace of official Calibre updates.

I think that your update will be seamless. Just make a backup "just in case".

When I am migrating to the new version of Linux distribution or to the new disk or a new PC, I just copy the Library directory and the ~/.config/calibre/ directory and install Calibre from the official site. I have never had a problem.
I have even migrated from the version of Calibre as packaged by a distribution to the official one in the past.

[haertig]

Originally Posted by kacir

The configuration data is stored at
~/.config/calibre/
Make a backup. This contains all plugins and configutations.
Configuration of custom columns and other library-related things are stored at the Calibre Library directory. I want to believe you already have a good backup of this directory. Make another one ;-)

Thanks. I figured Calibre would keep it's config stuff somewhere in my $HOME directory (which is of course backed up!) Calibres storage directory is off on a separate drive, but that's backed up as well. Being a long time sysadmin/network guru at work, I'm big on backups. At home I backup double - once to an external drive, and a second time to a remote backup server (my own server - I don't trust other people with my data). True, for these backups, that "remote server" is only a Raspberry Pi hooked up to a WD Red 4Tb drive, but that setup serves the purpose magnificently.

But you never know what's going to happen when a program was originally written for Windows, then ported to Linux (I don't know if Calibre is like this or not). Some of the Windows native stuff with a hacked port to Linux will store data in locations who-knows-where. FWIW, I certainly don't think Calibre is any kind of a "hacked port'. What a beautifully functional and feature-rich piece of software it is. So well implemented!

[kacir]

Originally Posted by haertig

But you never know what's going to happen when a program was originally written for Windows, then ported to Linux (I don't know if Calibre is like this or not).

The Calibre was originally written on Linux, for Linux, so that Kovid Goyal could access his Sony PRS-500 e-ink reader(*) on Linux computer. This was the first e-ink reader that was available to ordinary people outside Japan. The name of Calibre was originally Libprs500.

At this moment Calibre is still being developed on Linux and ported to Windows ;-) Oh ... and also to macOS ... (and FreeBSD, but not by Kovid Goyal I think)

(*) Sony, being typical Sony, created a device that needed a special software to access the reader, with drivers only being available for Windows. All subsequent e-ink readers from Sony and other manufacturers(**) function simply as a standard UMASS device when connected to PC

(**) with notable exception of a few manufacturers of e-ink readers based on Android that use mtp:// protocol, so that they can limit what you can do with your own files on your own device connected to your own PC. Arseholes!

[haertig]

Originally Posted by kacir

(**) with notable exception of a few manufacturers of e-ink readers based on Android that use mtp:// protocol, so that they can limit what you can do with your own files on your own device connected to your own PC. Arseholes!

I noticed that mtp stuff. It can be overcome by moving eBooks to the SD card (assuming your eReader has one), pulling the SD card out of the device, and mounting it separately. That would be a pain if you needed/wanted to do it frequently, but shows just how silly the manufacturers thought processes are.

Kind of like when the Keurig coffee maker people came out with their version 2 coffee makers that would reject any third party coffee pod you didn't buy directly from them. Until someone thought, "Why can't I just peel off the top of a recognized pod and scotch tape that to the coffee maker so it sees that instead of an unrecognized pod?" Well, you can do that easily, much to the manufacturers chagrin after they spent so much money/time trying to come up with "coffee DRM".