Speeding inode access
I have an unusual application which modifies multi-MB parts of large files and requires atime enabled for audit info. I am currently using ext4 with the journal on SSD. The data io spread over the array members nicely, but the inode io doesn't. Any thoughts on getting the inodes inode io faster?
My thought is to create a raid0 with a small chunk size over multiple drives, the existing raid6 with a large chunk size, and then concatenate them so the inodes are at the start in the fast portion. I can increase the number of inodes to take up the whole "fast" part and keep data in the large chunk portion. I am not crazy about using some brand-new type -f file system, but I'll listen to suggestions.
Signature · 4 lines
Bill Davidsen<[email protected]> We are not out of the woods yet, but we know the direction and have taken the first step. The steps are many, but finite in number, and if we persevere we will reach our destination. -me, 2010