Replies: 2 comments
|
Yes, we still have the limit, this is because we still use 32bit inodes, this is because converting to 64bits will cause a LOT of problems with 32bit OSes, which are still widely used. Snapshots create new inodes for files and directories, so yes, they will "use up" the limit. |
0 replies
|
Hello, I have a question. Why not create a separate mode that would use 64-bit inodes during cluster initialization to remove this limitation? Some physical servers have a large amount of RAM (several terabytes), meaning the metadata will fit in the master server's memory, even if there are more than 2 billion files in the cluster file system. |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Hello
I found this number in the doc, i am wondering if this number was still valid.
Nowaday, such number doesn't seems that high.
(My max is 1 billion inode used by a single group on a single server, its not ..yet... on moosefs)
I am wondering if 2Bn is the max number of inodes ( files+folder+link ) or files only.
If snapshot/archives, when used also decrease that count.
I didn t found this info in the doc, but is it possible to somehow increase this limit ?
All reactions