Manozco

That sounds like a nightmare...

I've extensively used pgreplay for work and I'd like to add a few things: - I recently (May 2018) added in pgreplay a dry-run mode where it just parses the log and print stats (no replay here). This could be used to benchmark the log parsing part. - Instead of parsing the log as you did, I would have configured PG to ouput logs in csv (it can even does both at the same time I think), would have been easier to parse - pgreplay has a binary replay file format well documented, I would have used that in your case probably

Otherwise seems like a nice tool, I might give it a try in the future :)

The Humble Bundle statistics reject your affirmation ;) http://cheesetalks.net/humble/

Linux users are paying more per bundle than other OSes users

You can monetize based on the time to read per person divided by the total time of this person. This way you eliminate {speed,slow}readers.

Fastmail has aliases if you use your own domain, and you can combine with + adressing: alias amazon@yourdns.com to username+amazon@fastmail.com and then all the emails are going into your folder amazon :)

I'm not seeing the usage here. What would need a (very very) performant temporary storage that you could not achieve with io volumes ?

I did not investigate MySQL code (or other product) but I would be very surprised if PostgreSQL is the only one impacted by this issue.

it depends, let's encrypt might be performing request from different sources (and then it would need a bigger BGP hijack to fake them)

Nah, you need to make them think it is going to bring them money. They'll happily take someone's else idea in this case ;)

Technically, certificates automatically validated only guarantees that you are on the website that let's encrypt thinks correspond to facebook.com. MiTM state wide could tamper with it

I did not see it on the forum, but seeing that the wildcard feature requires DNS-01 challenge for getting the certificates, does it mean automatic renewal is impossible without DNS api ? (or is it possible to renew without the dns challenge ? )

If your key is compromised, anyone can send that message and sign it with your key... :)

What's the meaning of COLA ?

Silence is a nice feature for SSD :)

SQLite

If you have really critical data, I expect you to run the latest possible kernel on part of your cluster. That allows you to see potential regressions and also to test your backup/recovery solution in case something that bad happens.

I cannot go into much details about what I do, but basically we have some kind of "network operators" that are the only one using the write db (and not a lot) and then we have a massive amount of computers doing reads on the system. We don't need citus for that, but PG replica does the job very well

NetBSD because "Of course it runs NetBSD" :)

You should call that Privacy Aware Gif-As-A-Service and monetize just that :D

I should definitely try that for replacing Insert with something else (probably delete)

I recently did my GPG setup. For the day to day operations, I have subkeys stored a yubikey. For the storage and backup, I use the nitrokey storage (https://www.nitrokey.com/), open hardware, open source usb key with hardware encryption. It's compatible with all os and there is a non encrypted partition (on which I put the app needed for main OSes) As a bonus, the nitrokey is also a smartcard openpgp (so I have a smartcard backup) and also has slots for 2-FA (So I have a backup of that) I plan to buy a second one that I would store outside of my home.

I remember reading some articles about the difficulty to have a fingerprint sensor under a screen, it might well have been easier to have FaceID (easier does not mean easy)

Well it's definitely possible to have me looking my phone held by someone else. And it's then too late...

I'm genuinely interested in knowing how apple can tell that FaceID is better than TouchID

- TouchID is already very fast

- I can give access to someone else with TouchID without giving my password

- It's unlikely that someone will be able to unlock my phone without me knowing it when using TouchID

- In case of coercion, I still have the possibility to give the wrong fingerprint 9 times before the good one

- I have to voluntary give my agreement with TouchID for an action (think apple pay)

All of that makes me think that they are trying to sell a feature that is only due to their engineering team unable to put TouchID on the Iphone X. By every real world metrics, TouchID is better in my opinion...

I'm not ultra familiar with the topic so if someone wants to correct me please do but :

- Linux has always been described as just a kernel, which translates as just a syscall table. The fact that this table is stable or not is not relevant here.

- *BSD on the other hand are shipping a kernel plus a lot of libraries/binaries, if you want to simulate a BSD system, you have to expose those libraries/binaries.

It's not so much a technical difference, it's more of a different approach to OS development (kernel space vs kernel/user space).

You might need that when you want to reimplement Linux, the Joyent team did that on their OS (derived from solaris) so that user can run linux binaries on a solaris kernel (so thay have dtrace, zfs, mdb, ...) Bryan Cantrill did a bunch of conferences on that (one here: https://youtu.be/TrfD3pC0VSs)

The idea behind is that Linux is only a list of syscalls, if you are able to reimplement them, you reimplement linux, you don't need anything else. On the contrary if you want to reimplement a BSD you need to reimplement their libc (and perhaps some other libraries)

Yes, sure, but executing the query a couple of times on the command line is also something I do

Yes, in many cases (until I stopped using it), it messed up my benchmarks. I was testing some queries, and often it was: - pgserver fast, returns 10000 rows, pgcli takes time to parse -> result: total time 10s - pgserver slow, returns 1 row, pgcli instant -> result: total time 10 sec.

As we are working in efficient C/C++, our internal time to parse the query set is closer to psql than pgcli, so we tend to prefer the pgserver fast, but the benchmark using pgcli did not clearly show the best time.

Also, when you are testing a query on one machine, you have to carefully monitor the CPU usage to differentiate the CPU pegging by postgres from the one by pgcli.

All those issues were gone when I got back to standard psql.

What will we answer in twenty years when our children will ask: "Daddy, where were you when they made the web a such shitty place? "

Then use Atom and don't comment The parent comment is valid for every software in the world, if you make your living with a 80$ software and you don't want to pay for it, you have big issues