1. 58
  1. 10

    This is an amazing example of technical investigation and writing. I think I found something noteworthy at each recursion of links I followed from the main story, e.g., the current mailing list software at lore.kernel.org.

    Also, it’s going to have me wondering about several bizarre rsync stalls from back in the day. Maybe that one time wasn’t caused by spiders in the server case in our unfinished basement garage lab. :)

    1. 3

      I’d be interested to compare with FreeBSD’s network stack. I wonder if the problem exists there, too.

      1. 2

        As is not unusual in these cases, we had more urgent systems and issues to attend to, so we labeled this a “race condition in rsync” that we should definitely look into at some point, and worked around it by throttling the rsync transfers.

        Until it started biting us every single day.

        Kudos for them to have persisted and looked for the Zebras! (It’s Zebras all the way down!)

        1. 2

          Great work investigating!

          1. 2

            As part of our standard toolkit, we provide each developer at Skroutz with a writable database snapshot against which she can develop.

            I understand that historically we used to use the masculine pronouns in gender-neutral statements but this “revenge equality” of using the feminine pronouns now is silly when we have a perfectly good “they” in English.

            What I mean… When I see “he” in such cases, I understand that it’s most likely just the old convention used unconsciously. When I see “they”, I acknowledge they use the modern convention. But when I see “she”, I see it as an unhealthy revenge for people using “he” for such a long time. Can’t we just switch to the neutral pronouns directly?

            Disclaimer: I’m not a native speaker of English and not a speaker of Greek at all, so it may be more of a cultural thing instead.

            1. 7

              I generally prefer neutral pronouns, but the idea that it’s a “revenge” seems pretty weird to me.

              More charitably, when you read ‘she’ and think ‘are they trying to make a point’, that’s a perfect opportunity to consider how others feel about the constant, far-more-widespread use of ‘he’ (Poe’s law - are people using ‘he’ making a point that mostly only men can code? Because I guarantee that’s a real opinion people express in private).

              1. 3

                But when I see “she”, I see it as an unhealthy revenge for people using “he” for such a long time.

                IMO this has nothing to do with the author and a lot to do with biases.

                Many books written 2015-onwards use he/she interchangeably in the examples. Very well known, respected practice. I try to avoid the conundrum by using a noun e.g. “the engineer <…>”.

              2. 1

                If the author(s) are reading this:

                About the Skroutz Engineering Blog Created with care by the Skroutz engineering team.

                I have no idea what Skroutz is and I don’t speek Greek. If you have a nice company tech blog, tell me more about your company! (Yes,the “We are hiring” link explains it a bit)

                1. 5

                  To be fair, the company’s services themselves are mostly aimed at the Greek market – Skroutz started out as (and still is) a price comparison website for Greek electronic storefronts, but has expanded to fulfilling orders and acting as a bona fide storefront for some companies that don’t wish to maintain their own. They’ve been a very good service whenever I’ve looked to buy something in Greece, and am glad to see this post make the rounds, as my impression is that you get very few of these sorts of engineering organizations in Greece.

                  Their development setups are very interesting as well – copy-on-write against a shared development database (which itself is based on anonymized production data) sounds like an excellent solution to a lot of issues: performance tests are honest and issues with runaway SQL queries are caught in dev, developers all see the same base data, so there’s less chance of mismatches, rollbacks are made much simpler, etc. I wonder how CoW works in terms of updating the base snapshot when changes to the schema have been made in a developer’s workspace.

                  1. 1

                    To be fair, the company’s services themselves are mostly aimed at the Greek market

                    Sure, but I specifically meant the /tech blog/ part, and in this case most people who aren’t in Greece probably don’t care about the company per se, but it’s still nice to know what the company does, to put their solutions into perspective. Doesn’t take away from the content in any way, but “we mostly do e-shopping” is already a pretty good context which I had to grab from the job ad.

                2. 0

                  Amusingly, I experienced this one at work.

                  The workaround was to timeout and resume (fortunately very doable with rsync).

                  Kudos for doing the sheepshaving. Just remember Linux has millions of LoCs and trying to fix it is a losing game.