TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Tmpfs inode corruption: introducing inode64

161 pointsby r4umover 3 years ago

8 comments

kzrdudeover 3 years ago
Is the mount option needed for backwards compat? Any system that creates that many files, should be large file aware, hopefully? Maybe it comes down to many files being created (and deleted) during the lifetime of the tmpfs mount, i.e usage time, not size, without any counter reset.<p>A system that uses &quot;our identifier space is approximately infinite&quot; can be fragile in a surprising way. A system that uses &quot;our identifier space is proportional to the physical size of the medium&quot; is more humanly predictable in its limitations.
评论 #28414534 未加载
jablover 3 years ago
Similarly, XFS has the inode32&#x2F;64 mount options. Since kernel 3.7 inode64 is the default: <a href="https:&#x2F;&#x2F;man7.org&#x2F;linux&#x2F;man-pages&#x2F;&#x2F;man5&#x2F;xfs.5.html" rel="nofollow">https:&#x2F;&#x2F;man7.org&#x2F;linux&#x2F;man-pages&#x2F;&#x2F;man5&#x2F;xfs.5.html</a>
评论 #28413369 未加载
smitty1eover 3 years ago
I have not read an article in quite some time that takes a complex topic and lays it out so clearly.<p>Thank you!
opkover 3 years ago
I enabled `pam_namespace` for per-user `&#x2F;tmp` and `&#x2F;var&#x2F;tmp` on the systems I administer with those filesystems being tmpfs and xfs respectively. We get odd problems weekly where one or other user can&#x27;t write to `&#x2F;tmp` (but nsenter fails to reproduce it) but have never had issues with the `&#x2F;var&#x2F;tmp`. We came from Solaris where tmpfs was standard for `&#x2F;tmp` and it surprised me to find that it seems to be an unusual choice on Linux. Are there reasons it isn&#x27;t widely deployed on Linux? I can&#x27;t think why inode reuse could cause the problems I experience but if I can enable this inode64 option, I&#x27;ll give it a try.
评论 #28417188 未加载
tyingqover 3 years ago
I&#x27;m somewhat intrigued that this problem has remained for decades after 64 bit inodes were available in Linux.
opkover 3 years ago
I&#x27;ve not come across FS_IOC_GETVERSION before. Should the many user-mode applications that rely on st_dev, st_ino being enough to identify an inode also be calling this ioctl? Or is it something you only need to worry about in limited circumstances?
cpcallenover 3 years ago
Can someone explain how tmpfs &#x2F; &#x2F;dev&#x2F;shm track files if not by inode number?
评论 #28413432 未加载
评论 #28412849 未加载
评论 #28412820 未加载
评论 #28412917 未加载
ameliusover 3 years ago
While we&#x27;re at it, can we also increase the default max number of hardlinks on ext4 filesystems? And the max number of commandline arguments?