Talk:inode

"The kaki of a file system that makes use of the concept of inodes surprises many users who are not used to it at first:" What does "kaki" mean? 213.94.245.1 19:50, 27 November 2005 (UTC)Reply

Hehe, obviously vandalism... Reverted. 70.82.141.92 01:08, 4 December 2005 (UTC).....Reply

I elaborated on the formal definition, and grounded it with some references. jdmartin86 —Preceding undated comment added 04:50, 25 November 2013 (UTC)Reply

Where is the inode information stored? With the file? Or in a special directory? — Preceding unsigned comment added by 65.189.137.114 (talk) 06:58, 20 March 2006 (UTC)Reply

Check this out -> Unix_File_System — Preceding unsigned comment added by 200.45.225.22 (talk) 16:04, 23 March 2006 (UTC)Reply

It depends on the file system type. Guy Harris (talk) 21:42, 31 May 2023 (UTC)Reply

For anyone who knows enough about it to do so, some discussion here about the balance between having enough inodes available and using up too much space when building a filesystem would be helpful. Examples might include disk space usage and inode count for the same number of files when different -i values are used with mke2fs. --joshua orvis 14:49, 8 January 2007 (UTC)Reply

The whole section seems to be a bit schizophrenic. For example:

1. inode file systems can fragment as severly[sic] as FAT based file systems, yet universally inode file systems lack deframenation tools.
2. An inode file system would have to be offline to be fully defragmented on most systems -- but some online defragmenation tools exist.

The subsection about i_generation seem to really be either NFS only topics or NFS and Linux only topics.

There's also some odd comparison to FAT32 and NTFS.

I submit that this whole section should be removed. IMHO it has, at best, dubious value. --J3gum (talk) 18:04, 21 February 2008 (UTC)Reply

The first paragraph here appears to contradict itself: first saying that the number of inodes (and therefore max files) is fixed, but then saying that most file systems can handle unlimited files —Preceding unsigned comment added by 82.11.19.16 (talk) 00:08, 12 March 2008 (UTC)Reply

I agree. I am removing the later phrase, as it is probably the incorrect one. SSPecter ^Talk|E-Mail ◆ _{12:08, 19 March 2008 (UTC)}.

Many, not all, file system types fix the number of inodes at creation time. I've fixed the text. HughesJohn (talk) 10:36, 24 September 2008 (UTC)Reply

In the following sentence, it is unclear which graph has N-1 edges: "This made the directory structure into an arbitrary directed graph as opposed to a directed acyclic graph (DAG), a connected graph with N-1 edges for N nodes." --Bobbozzo (talk) 21:09, 15 December 2009 (UTC)Reply

"This made the directory structure into an arbitrary directed graph as opposed to a directed acyclic graph (DAG), a connected graph with N-1 edges for N nodes."

Neither of these classes of graphs (arbit. directed or DAG) necessarily have N-1 edges for N nodes. In fact the first diagrams of the respective links in the text show counterexamples. |E| = |V-1| is the characterisation of a tree, if I remember correctly. Also this part of the text seems only marginally to do with inodes. Chaos.squirrel (talk) 13:45, 14 June 2010 (UTC)Reply

I agree with Chaos.squirrel. A connected undirected graph without cycles does have N-1 edges for N nodes and is thus a tree. However, with DIRECTED graphs, two ways of getting from one node to another don't necessarily imply existence of a cycle. As it seems that a long time has passed since Chaos.squirrel's comment and noone answered to it, I will take the liberty and remove the invalid statement about DAGs myself. Czestmyr (talk) 11:15, 5 September 2010 (UTC)Reply

Page 57 of the The UNIX Programming Environment states "The administrative information is stored in the inode (over the years, the hyphen fell out of "i-node"), [...]" implying that the i stands for information. Should this be added? --BiT (talk) 07:18, 22 November 2010 (UTC)Reply

On page 95 of Jerome H. Saltzer's book Principles of Computer System Design, he writes:

To support this requirement, the UNIX file system creates an index node, or inode for short, as a container for meta-data about the file.

I'm not sure how authoritative this is, though. Clark3934 (talk) 03:39, 24 February 2011 (UTC)Reply

The most authoritative source I can think of is the original paper on UNIX: The UNIX Time-Sharing System. In Section 4, "Implementation of the File System," Thompson / Ritchie write: "This pointer is an integer called the i-number (for index number) of the file. ... The entry thereby found (the file's i-node) contains the description of the file as follows."

This combined with the bit from The UNIX Programming Environment seems to make "index node" the correct definition. Apologies for the shoddy reference formatting. Arek (talk) 21:27, 21 March 2014 (UTC)Reply

How to retrieve it ? ( df -i <device> ) How to change it  ? ( tune2fs -I <device> ) where to look the description of this procedures? — Preceding unsigned comment added by Arsen.Shnurkov (talk • contribs) 08:53, 13 May 2012 (UTC)Reply

please use the explanation and pic from the german wiki -- 80.245.147.81 (talk) 06:57, 13 March 2013 (UTC)Reply

Someone boldly revised all the occurrences in this article of 'inode' to 'i-node'. While this might be appropriate, it probably should have had some discussion or evidence that is the majority usage (per WP:COMMONNAME). As the editor was not registered, s/he can be excused from not also renaming the article.

As a starting point, a Google search of Unix inode indicates 696,000 occurrences while Unix i-node has 19,200,000. Linux inode has 1,860,000 and Linux i-node has 9,300,000. At first blush, this seems to weigh heavily in favor of i-node.

However, inspecting the results shows that Google appears to have ignored the hyphen as the first hyphen results do not show any context with a hyphenated word. Therefore I repeated the searches with quotes around i-node. For this Linux "i-node" returns 89,700 results and Unix "i-node" returns 81,600 results. This strongly indicates that inode is the most common. —EncMstr (talk) 21:47, 24 November 2013 (UTC)Reply

Just another argument in favor of inode vs. i-node: Google Ngram Viewer shows that inode has been more common in books since the early 80s. I say change the article back to using inode. Now I'm really curious about what inode meant in the 1800s, though! Arek (talk) 21:19, 21 March 2014 (UTC)Reply

Hello fellow Wikipedians,

I have just added archive links to one external link on Inode. Please take a moment to review my edit. If necessary, add {{cbignore}} after the link to keep me from modifying it. Alternatively, you can add {{nobots|deny=InternetArchiveBot}} to keep me off the page altogether. I made the following changes:

• Added archive http://web.archive.org/web/20160211033517/http://www.ibm.com/developerworks/linux/library/l-linux-filesystem/ to http://www.ibm.com/developerworks/linux/library/l-linux-filesystem/

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at {{Sourcecheck}}).

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 5 June 2024).

• If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
• If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—^{cyberbot II}_{Talk to my owner:Online} 12:59, 28 February 2016 (UTC)Reply

It has been noted at Stack Exchange that this article erroneously talks about a "POSIX i-node description" when in fact there is no such thing in the standard. Stack Exchange notes that, obviously, on-disc i-nodes have no need of a device ID. Furthermore: in on-disc i-nodes, the i-node number is not within the i-node itself, it being simply the index into the i-node array after all. Nor, indeed, does the POSIX stat structure, with which i-node has been erroneously conflated here, contain any pointers to disc blocks.

The "POSIX description" is not POSIX, and is a confused mixture of in-memory i-nodes, on-disc i-nodes, and struct stat. All three of these are different things; and only the last of them has anything at all to do with POSIX.

The article does not even explain the in-core/on-disc distinction, by the way. Jonathan de Boyne Pollard (talk) 07:20, 19 August 2017 (UTC)Reply

Thanks Jonathan, I have added a {{Disputed section}} redirecting here. — Xavier, 16:12, 20 August 2017 (UTC)Reply

@Xavier and Jonathan de Boyne Pollard: I clarified that POSIX calls them file serial number, and reordered attributes as defined by the standard. Further explanations may be still needed, for example in memory vs. on disc. Yet, since what is disputed is whether POSIX defines inodes, I think said clarification suffices to remove the Disputed section template, so I'm going to do it. ale (talk) 16:00, 15 January 2018 (UTC)Reply

Thank you ale. I'm fine with removing the banner. I have slightly changed the section, following your edits. First, the second device ID is not the device ID of the underlying FS but a device ID that is only meaningful when the inode is associated to a special file (this is usually the combination of the major and minor device numbers on the majority of Unices). Second, I made it clearer that the list is actually the data returned by the stat system call, and not attributes a file "must have". — Xavier, 18:47, 15 January 2018 (UTC)Reply

The link in JVSANTEN reference did not take me to this website: https://www.linux-faqs.info/general/difference-between-mtime-ctime-and-atime

Instead, it took me to this website: https://northgatevineyard.mobi/lk21/ 2600:1700:4FD0:D70:4817:4AD6:EC57:2B02 (talk) 05:20, 24 April 2023 (UTC)Reply

Domain names sometimes don't get renewed and get grabbed by other people. Fortunately, the Wayback Machine preserves archives of a lot of sites, including that one, so the reference has been fixed to use an archived copy (and to mark the original as "usurped"). Guy Harris (talk) 06:33, 24 April 2023 (UTC)Reply