A reader writes:
So, I’m partway through your book, and I want to check on hard links. When you say it creates another door to the same body of data, is that literally true? For example, if all I have in my ./Documents directory is 001.doc (size 512K), and I hard link as hardlink.doc then run ls, do I see A or B?
A) 001.doc 512K
B) 001.doc 512K
If it’s not B, I’m not sure what the point is, other than maybe not changing the timestamp.
An interesting question! The main point of hardlinks is to allow a single data entity to be shared by multiple entries in the file system. If I saved this posting as “001.html” on my server, then created a hard link to the same file from “002.html”, then I could subsequently delete “001.html” and “002.html” would be unaffected, with the same contents within. It’s only when all links to a data entity are removed that the data gets hooked into the free list and is essentially recycled for later reuse.
In terms of your specific question, the system will view the two files as having separate content. This can be demonstrated thusly:
$ du 249 . $ ls -l total 248 -rw-r--r-- 1 dtint vuser 247034 Aug 24 03:24 some.data Now let's create some hard links. Notice "ls" reports 248 blocks in this directory $ ln some.data link2 $ ln some.data link3 $ ln some.data link4 $ ls -l total 992 -rw-r--r-- 4 dtint vuser 247034 Aug 24 03:24 link2 -rw-r--r-- 4 dtint vuser 247034 Aug 24 03:24 link3 -rw-r--r-- 4 dtint vuser 247034 Aug 24 03:24 link4 -rw-r--r-- 4 dtint vuser 247034 Aug 24 03:24 some.data As expected, "ls" now reports 992 blocks in use, seeing each of the new hard linked files as new data. But... notice what "du" reports, since it actually looks deeper into the file system: $ du 249 .
As you can see, some elements of the file system see hard links as new and separate data entities (e.g. “ls”), while others dig further into the file system to ascertain actual data blocks used and associated with the current location (e.g., “du”). This makes sense because otherwise if you had lots of hard links you could theoretically end up in a situation where you have greater than 100% disk space utilization, which would clearly be a bit confusing at best!