Jubarb, Inodes, And Always Biting Off More Than I Can Chew

Herb’s Blog, Herbdate 23446 – 1204

Here’s the haps:

Being self-hosted is not without its perils. I have been getting notices that my site has been down for extended periods, some people having more problems commenting, and my storage fluctuates when I run backups. People who are hosted on WordPress.com have restrictions and limitations they have to follow but people like me, whom I call “Dot Orggers,” have different problems and we don’t have access to Happiness Engineers but rather have to search the forums of WordPress.org and other sources to figure things out. I imagine it comes out to six of one and a half a dozen of the other when you compare the two. I like the independence of being self-hosted although there can come pitfalls with it. I have been learning things and so there is an upshot to my adventures.

I have inode problems. While that may sound like something a cream should be able to fix, it actually has to do with the way files are stored on Linux systems. Most web hosts offer either a Linux-based server or a Windows server when you first sign up. Linux has a lot going for it in running websites and since I first uploaded my main site, herbthiel.com, a couple of years before I started this blog, it has served me well. When I decided to leave Blogdrive and make my own blog I tried several different types of blogging software and finally ended up using WordPress. Drupal, Joomla, the now-defunct Nucleus, et al. all have their pros and cons but one WordPress advantage is the Jetpack plugin which allows me to connect with the community of dot commers, which many of you, whom I have found by serendipity, are.

But, back to Inodes (I knowed he was going to come back to it, I just knowed it.). Since WordPress has switched over to this Block Editor I have had increasing problems with the limits of my disk storage on my host’s server. The file systems of Windows and Linux operate in very different ways. In Windows, what is called the metadata for a file, e.g., its name, creation date, time stamp, and other information about a file is stored in a directory with the data of the file itself. In Linux, the metadata is stored separately from the data. The metadata includes a number assigned to the data itself. This metadata is stored in inodes. The directory information is also stored in an Inode, so no matter the size of the file, it generates the same inodes for the metadata. It’s possible, especially if you have a large number of small files, to “fill up” the hard disk and not use all the space available on that disk. This is because of a limitation on the inodes themselves called bytes-per-inode. If you have a very small file with a very long name you could use more inodes than a large file with a short name. My host puts it like this, “The longer your filenames, the fewer files you can have in a directory (file name lengths of 1-16 characters count as 1 inode, 17-32 characters count as 2, 33-48 characters count as 3, and so on).”

I apologize that this explanation is clear as mud but I am just now starting to get to know some of this stuff. One of the better explanations I have read on one of the forums was that Linux is like an old-fashioned library. The book is the data, the inode is the Dewey Decimal number, and the card catalog is the directory. Just as there can be twenty cards in the catalog for one book, many directories can point to the same file. I bog down here because the way I understand is that each directory has its own Inode. The folder in the WP-Includes for the blocks contains 785+ entries alone, which are considered part of the WP core and cannot be removed. Now most of the forums I looked at all made it sound like WP was not the real culprit and I should look for other possible sources for the spikes in my Inodes. Which is probably kind of true, but not completely.

When I have uploaded images, especially memes, they have very long file names that look like the digital version of alphabet soup. My own images often have long, descriptive file names. A lot of the meme file names (and many of my own) have over 40 characters. That’s bad enough but then when you upload a file to WP it creates a minimum of 3 different versions of the same file in different sizes. Thus this meme:

which has the file name 01FQ1REKRSGTT6H7PYJEGWR5VE.jpeg (133.68 kb) also becomes 01FQ1REKRSGTT6H7PYJEGWR5VE-150×150.jpeg (7.65kb) and 01FQ1REKRSGTT6H7PYJEGWR5VE-300×300.jpeg (21.68kb). The files themselves are really pretty small but this generates a gob of Inodes. Depending on what theme you have, it could generate considerably more copies with very little difference in size. I read one post where a theme had created 13 copies of each file for a person. This is really silly to me. That the program can’t simply scale the image is beyond my ken. Anyway, I will be making some changes to the blog but I expect that getting rid of and/or redoing some of the meme dumps is probably where I will start. Then renaming files I need/want to keep.

---

I think in the future, If I try to catch up on the A – Z theme, I am going to dump obsolete words and pick fresh ones because this was a mild nuisance. The word generator was set to use the entire dictionary instead of just common words, but still…

The term Jubarb, which has not been used since the 1700s, means a succulent houseplant, now called a houseleek, of which there are 40 species. Now, when my house leaks, I call either the roofer or the plumber. According to a medical publication on the NIH.gov website called (in part), Ethnopharmacological uses of Sempervivum tectorum L. in southern Serbia:, it’s good for some earaches.

A weasel walks into a bar and the bartender goes, “Wow! I’ve never had a weasel in here before! What’ll you have?”
“Pop,” goes the weasel.