Skip to main content


Hilarious bit about NetApp ...

『A completely different way of thinking can be found in Log Structured File Systems , first seen in Ousterhout’s Sprite, in the early 1990. But performancewise the original implementation that was a failure: the first successful and performant implementation can be found - made by Oursterhout’s students - in Solaris ZFS, and at the same time in NetApps WAFL, which promptly engaged in a legal battle over patents. In parallel, many ZFS ideas are being reimplemented about half a decade later in Btrfs ...』

... from ...

What are the problems with #POSIX? 20201005,
by Kristian K,
blog.koehntopp.info/2020/10/05…

#unix #fileSystem

This entry was edited (1 year ago)
in reply to parvXtl

I think this part of the reported history is a bit wrong. While the announcement seems to indicate @bcantrill was involved, my days at Joyent made me believe that Mark Cavage and @dap were the driving forces behind Manta. From Mark’s telling in 2013 (web.archive.org/web/2013062900…) Jerry Jelinek wrote the first code and others pitched in on other things.

LXC was around in 2008 and I think virtuozzo was earlier than that. Saying zones and jails were the only containers in 2013 is not right.

in reply to Mike Gerdts

Re: Manta, I read that as Bryan C being credited for the idea behind it, in form of an inspiration or a prediction. Is that impression incorrect?

@bcantrill @dap

This entry was edited (1 year ago)
in reply to parvXtl

@bcantrill @dap I had previously heard it described as something Mark invented without a lot of participation from others. The announcement seems to spread credit around with at least some involvement from Bryan. When I was at Joyent, almost every new feature would have some involvement from Bryan and several others. It’s too easy to point at the most famous person in the room where it happened and assign all the credit. Maybe it’s due. I genuinely don’t know.