karl@grebyn.com (Karl Nyberg) (08/02/89)
I need more inodes for my news spool partition. According to my manual page for newfs (8) on paper, you use the: -i number of bytes per inode This specifies the density of inodes in the file system. The default is to create an inode for each 2048 bytes of data space. If fewer inodes are desired, a larger number should be used; to create more inodes a smaller number should be given. Well, increasing the number does indeed decrease the number of inodes. However, decreasing the number, all the way down to 2, has absolutely no effect whatsoever. The system continues to make one inode per each 2048 bytes. Am I missing something here? Should I really be using mkfs (which is getting passed the correct parameters, using the -v flag to newfs)? -- Karl --
grr@cbmvax.UUCP (George Robbins) (08/03/89)
In article <12244@grebyn.com> karl@grebyn.com (Karl Nyberg) writes: > > I need more inodes for my news spool partition. According to my manual page > for newfs (8) on paper, you use the: > > -i number of bytes per inode > > Well, increasing the number does indeed decrease the number of inodes. > However, decreasing the number, all the way down to 2, has absolutely no > effect whatsoever. The system continues to make one inode per each 2048 > bytes. Am I missing something here? Should I really be using mkfs (which > is getting passed the correct parameters, using the -v flag to newfs)? There are interdependencies between various of the mkfs parameters such that mkfs silently changes the supplied parameters when a limit is execeeded. In particular, there is a limit on the maximum number of inodes per cylinder group. Overriding the size of the cylinder groups to make them smaller will allow you to get more inodes. I seem to recall that changing block/frag size parameter will also do this, though perhaps indirectly by affecting the cylinder group size. You can verify what you are actually getting by using the dumpfs(8) utility. See /usr/include/ufs/fs.h for some factoids and the section on the Berkeley Fast File System in the Supplementary Documentation for added confusion. -- George Robbins - now working for, uucp: {uunet|pyramid|rutgers}!cbmvax!grr but no way officially representing arpa: cbmvax!grr@uunet.uu.net Commodore, Engineering Department fone: 215-431-9255 (only by moonlite)
chris@mimsy.UUCP (Chris Torek) (08/04/89)
In article <12244@grebyn.com> karl@grebyn.com (Karl Nyberg) writes: >I need more inodes for my news spool partition. ... [newfs -i option] >However, decreasing the number, all the way down to 2, has absolutely no >effect whatsoever. The system continues to make one inode per each 2048 >bytes. I wrote a nice article explaining this problem in response to a message asking about newfs on Suns, and then forgot to save it. The short answer is `fixed in 4.3BSD-tahoe; beat on DEC to update their ufs code; in the meantime, try -c'. The long answer (rather shortened) is that there is a hardwired limit of 2048 inodes per cylinder group. The only way to get around this in the current system is to make smaller cylinder groups. This is not always possible, however, as the rotational layout code depends on the number of `cylinders per cycle' (essentially, how soon does the rotational layout pattern repeat) and you cannot lower the number of cylinders per group below fs_cpc. fs_cpc is aggravated by disks with an odd number of sectors---the closer the number of sectors to a power of two, the lower fs_cpc will be. In particular, some drives with 67 sectors per cylinder cannot have less than 8 cylinders per group. -- In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7163) Domain: chris@mimsy.umd.edu Path: uunet!mimsy!chris
chris@mimsy.UUCP (Chris Torek) (08/04/89)
In article <18890@mimsy.UUCP> I wrote: >In particular, some drives with 67 sectors per cylinder Er, make that `sectors per track'... -- In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7163) Domain: chris@mimsy.umd.edu Path: uunet!mimsy!chris