My configs and locations?

[devtech8]

So am running my Calibre on Ubuntu and my drives are split up so I like to keep my library on mounted raid drives. I just let Calibre install in default location, but am guessing that means the cache and virtual library, configs, and any other items will grow larger on this smaller boot drive.

I understand about the environment variables per the documentation, but cannot figure how to set or change them. If I simply run

CALIBRE_CONFIG_DIRECTORY

, it does nothing. Help on this please from doing so via command line?

Additionally, I do have a client install on a Windows machine which also see thew library and all from the GUI, can I also adjust from there and if so, how?

Thanks in advance!

[BetterRed]

Originally Posted by devtech8

So am running my Calibre on Ubuntu and my drives are split up so I like to keep my library on mounted raid drives. I just let Calibre install in default location, but am guessing that means the cache and virtual library, configs, and any other items will grow larger on this smaller boot drive.

I understand about the environment variables per the documentation, but cannot figure how to set or change them. If I simply run

CALIBRE_CONFIG_DIRECTORY

, it does nothing. Help on this please from doing so via command line?

Additionally, I do have a client install on a Windows machine which also see thew library and all from the GUI, can I also adjust from there and if so, how?

Thanks in advance!

CALIBRE_CONFIG_DIRECTORY is an environment variable not a program you can run! You must set it, using operating system specific facilities, to the path where your calibre configuration data is located:

Search The Internet for "how to set environment variables in <your operating system>"

To find out where your config data is, click the Preferences->Miscellaneous->Open... button.

BR

[devtech8]

In Ubuntu, is it referring to this:

sudo nano /lib/systemd/system/calibre.service

for setting the environment variables?

So then

CALIBRE_CONFIG_DIRECTORY=path/path/xxxx

?

[devtech8]

Additionally, when I run this:

mark@morpheus:/$ echo "$CALIBRE_CONFIG_DIRECTORY"

mark@morpheus:/$ echo "$CALIBRE_CACHE_DIRECTORY"

It does return any value. Is no value assigned by default?

[BetterRed]

I suggest you ask how to set Ubuntu environment variables at the Ask Ubuntu forum, 97.8% of calibre users use Windows or MacOS.

The default location is set internally to a standard location, the environment variable only needs to be set if you move your configuration elsewhere.

BR

[Adoby]

Linux/Ubuntu.

You set environment variables in either ~/.profile or /etc/profile or possibly in a script used to launch calibre.

If you look you can see how other variables are set there.

However it will NOT help in keeping stuff from the cache or significantly reduce use of the root filesystem. So don't do it!

All settings and data specific to one library is stored together with the library, except the name and location of the library.

Calibre will use /tmp and the cache in your home folder, but that will have to be fixed some other way. For instance by placing /home on another drive and adding an SSD for /tmp.

There is another environment variable for setting what path calibre should use for temporary files. CALIBRE_TEMP_DIR. It might be useful to change if you are in the habit of running a lot of batch conversions.

It is general good practice to separate the root file system and /home. Use different partitions. It is also good practice to have /tmp on your fastest drive, or even in RAM using tmpfs if you have plenty of RAM installed. That also may speed up some operations in calibre.

[devtech8]

I tend to only run WD Red drives for my storage on my Linux box.

[Adoby]

WD Red are fine drives. But make sure you also have good backups. And make sure that you know exactly what to do when one drive drop from the array. With modern big drives raid may no longer be a very good idea for personal use.

I use versioned full rsync backups and no raid for my main NAS storage, to separate backup storage servers. Actually, it is just a bunch of small single board computers, each with a really big HDD, running OpenMediaVault. But I think of it as my personal storage cluster. Sounds better.

[devtech8]

Originally Posted by Adoby

I use versioned full rsync backups and no raid for my main NAS storage, to separate backup storage servers. Actually, it is just a bunch of small single board computers, each with a really big HDD, running OpenMediaVault. But I think of it as my personal storage cluster. Sounds better.

But that does not reduce the number or size of drives you need. I figure folks use things like OMV cause they cannot get a grip on other distros. All in all, it works the same, but can get hairy depending on how complicating one makes it.

[Adoby]

Sure, OMV makes it easier to get a great headless fileserver running quickly. It is a full Debian distro, optimized for filesharing with OMV on top. I want my storage simple and interchangeable.

Make sure that you understand the difference between what raid provides and what backups provides. And that you have the skills needed to recover from failures. Even if you use raid you still need backups. And make sure you understand why raid with big drives may be a chain sequence disaster waiting to happen. If you need to rebuild a big array there are so much data that need to be shuffled around that it may cause more drives to fail in the process. With total loss of all data in the array as consequence. The only way to protect against that is with good backups.

So, no. With raid you need more drives. For redundancy AND for backup.

I have access to my files from the original single drive (12TB Ironwolf drives) OMV storage servers. If something goes wrong with them I have access to my files from my OMV backup servers, in several versions. And then I have offline backups on top of that.

What happens if your array becomes unsynced and you need to rebuild it? Can you access your files then? Only if you also have online backups.

[devtech8]

Ah. Yeah, I am not network admin or anything like that. I consider myself pretty technically savvy, but had not thought of storage that way. I know this is getting off topic, but how did you come to deciding as this? Am guessing to that this means you have several physical boxes running too? Any posts, etc... you can link to me to explain your setup and why?

[Adoby]

I had not a lot of space left on my old 12TB Synology 411j NAS. And I didn't want to put in even bigger drives. And it was slow.

So I started looking and comparing alternatives. Odroid HC2 was pretty new then so I bought one and a big 12TB Ironwolf drive to test. After some looking I decided on OMV as the dist to use.

I was very pleased with the performance, it can easily saturate a GbE network connection and it is pretty cost effective as well. At least if you calculate price per TB using the biggest NAS HDD you can find. The cost of the Odroid HC2 is dwarfed in comparison to the price of the HDD then...

The Odroid HC2 is a very tiny single board computer. About the size of a matchbook. 2 GB DDR3 RAM, USB 2, GbE and one SATA 3 port. ARM. And 8 cores. It comes with a big aluminum profile that is used for cooling and for mounting and protecting a HDD. Also the aluminum profile can be used to stack several HC2 together. That gave ideas...

So I ordered four more HC2s directly from Korea. Two are currently used as my "front line" storage nas. Two are used as my "behind the line" backup nas. One, with a 512 GB SSD is used as an app server where I run things like Emby, LMS and NodeRed and I also use it for system backups of the other servers.

All the servers are connected together in a GbE switch. And I access them over a fast wifi mesh, Asus Lyra.

If I need to I can easily move HDDs between HC2s or even to my desktop PC. And access the contents. So I got HW backup as well.

If I ever need more storage I can just add another HC2. And another. Scales well... There are 14TB Ironwolf HDDs available now.

I considered using some distributed filesystem like glusterfs, but decided not to. At least for now. I prefer to keep it simple. With more HC2s that may change, then glusterfs might simplify things.

The old Synology 411j is mostly turned off and used as offline backup.

All clients are either Linux or Android. I also use the storage servers for backups of my clients.