i read the paper.
i can believe there are still bugs, truth be told there are bugs in most software. all i can say is in my experience i have not hit any since the ephemeral snapshot fix. Honestly fosdil has been solid for me; i hope i have not just jinx’ed it :-) -Steve > On 15 May 2024, at 11:18 pm, hiro <23h...@gmail.com> wrote: > > noam said there's also some well-known issues with locks that still > keeps happening in fossil. did you watch the outcome of last iwp9? he > has the best description of the state of venti and fossil to date. > >> On Thu, May 16, 2024 at 12:08 AM Steve Simon <st...@quintile.net> wrote: >> >> >> Just a little first hand experience - I have run a fossil and venti server >> for twenty years now. >> >> Fossil suffered three problems in my opinion: >> >> Firstly it was not well publicised that fossil was never designed to cope >> with overflow; the sad truth is it fails catastrophically, as whitenessed by >> Brucee. it should cope better really, but when used as a write buffer for >> venti this is unlikely, it has never happened to me. >> >> Secondly it had a race condition in the ephemeral snapshot code. the only >> way to make fossil reliable was to turn these off. This bug was fixed by >> Charles (i think, though perhaps it was Richard) about 10 years ago and with >> the fix my server has been taking ephemeral snapshots ever since. >> >> Thirdly venti and fossil are slow. >> >> I think its sad (an emotion, sorry) that fossil sources where deleted from >> 9front but i do understand the principal of not distributing things the >> community does not wish to support. >> >> -Steve >> ------------------------------------------ 9fans: 9fans Permalink: https://9fans.topicbox.com/groups/9fans/Tad3dc0c93039a7d2-M0d5250f5084dfe501f2a0e2d Delivery options: https://9fans.topicbox.com/groups/9fans/subscription