mioelnir

Think flexible-sized partitions. Then forget it again because it is too wrong.

The one really nice scenario this opens up that is discussed in the pull request linked in another comment:

  1. Server A has an encrypted dataset in pool foo, currently not decrypted
  2. A can send full or incremental streams from that dataset to server B
     without decrypting
  3. B can receive those streams and import them as encrypted
     datasets into the pool without decrypting or really ever having
     even seen the keys
  4. Server A can restore from B as required
  5. The owner of the key material can log into B and unlock
     the dataset as if it were on A

This is a really nice, accessible way for encrypted remote backups.

2 minutes is cute. If you stress a CMS setup hard enough that the young generation is completely full, it will allocate directly in the old generation. This of course screws the full gc heuristic totally, up to the point where the GC is started too late and you fully run out of memory. At which point the JVM drops down to a single threaded oldschool serial GC as last line of defense. On a 96GiB heap, that thing can take hours; all stuck 100% on a single cpu with even signal handling suspended. Fun times.

That said, for heaps above 32ish GiB, we still go with our tuned CMS settings and overcommit one or two additional memory modules. It's a lot cheaper than the time it takes trying to tune in G1 on a large heap with a lot of gc pressure.

Unencrypted Netflix pushes 80~90+ GBit/s from their most recent OpenConnect revisions.

[1] https://twitter.com/ed_maste/status/655120086248763396

[2] https://twitter.com/scott4long/status/656219076629368832

[3] https://twitter.com/facepalm_tar_gz/status/71066891267375104...

[4] https://media.netflix.com/en/company-blog/how-netflix-works-...

But a hit to german trading to the UK is a small price to pay instead of the hit it would take by setting precedent that you can just willy-nilly leave with all the perks. And I can tell you, the german-language media is already full of "priority number one if to stop a cascading copycat effect".

As a snarky side remark, what will the UK do in the future instead of continuing to import cars, regardless of terms? Rely on its domestic automotive industry?

There is a saying I can't fully remember about the nobility of an old man planting a tree, knowing he will never sit in its shade.

At which point should you switch from trying to shape your generation's future to trying to shape the next one's?

As a german, there is a reason Hitler create a centralized government tuned for efficiency; and there is a reason Germany was rebuilt as an inefficient bureaucracy. Because in the latter you only get to move fast on something if every single involved person has a natural understanding that this is the right thing to do, has to be done, and why did it take so long?

Inefficient government makes it harder to move against your own population without open military action.

No, we all agree that, both historically and presently, our political leaders are incompetent and not to be trusted with such a sharp blade. They get children's scissors and should be happy about it.

1000 years of religious oppression.

> Unless of course UK retains the free market with EU.

Why would they?

> for the differing peoples of Britain and the Continent

So you on the islands are special but we monkeys on the continent are all the same? The UK has been half-in/half-out of the EU for the last 40 years, always wanting extra on all the benefits and never taking a downside in return.

I'm glad you've finally made up your mind, as one way or the other this shitshow will finally end.

The EU forces a minimum standard for you have to implement, correct. Like not pumping unfiltered sewage into a freshwater reservoir, giving a damn about air pollution, non-toxic children's toys etc.

These are not made-up examples.

Why would the EU give such a favorable deal to a country as unimportant as UK? The EU-member UK leaving is an important signal; but non-EU-member Uk is an irrelevant tiny market.

The Brexit camps idea that they could leave so they could finally pick whatever they like, as they have been trying to do for the last 40 years, is cute but laughably naive. They will not continue to receive the free travel, trade access and EU subsidies to the tune of 200 million euro per week. All of that will be revoked, only if to set an example that you will not get those benefits without buy-in.

In my view, there are currently two equally likely scenarios to unfold: they leave with nothing as farewell gift (as has already been hinted at by the EU), or they will ignore the vote and stay in because they realize how bad of an idea this was - but are now in such a bad position they will likely lose all special privileges they enjoyed over the last 40 years.

This can be clearly seen by how many leave-voters are now in shock that they leave, and want to be able to vote again to remain. This was a powerplay for a better negotiation position. It backfired and now they can eat it.

SSDs use CRC to deliver their probabilistic data storage approximation. FS level checksums are, among others, for when that fails.

FS level checksums also cover bitflips on your bus during the write phase (the drive will crc the already bad data) or during the read phase (you receive different data than your drive sent).

Between the ZFS-like features and the advertised "novel copy-on-write metadata" scheme, I would not be surprised if it was partially based on DragonFly's HAMMER/HAMMER2 filesystem.

I've ran the initial FreeBSD patchsets for ZFS support on a dual-p3 with 1.5g ram, so 1g for a recent version should be more than doable. ZFS on FreeBSD has become a lot better with low-memory situations.

There are two additional things to consider.

ZFS uses RAM mostly for aggressive caching to cover over both spinning disks and the iops tradeoff vdevs make over traditional raid arrays. Thus low memory is not such a big deal if you have a pool with a single SSD or NVMe device.

The other point to consider is that on at least any non-Solaris derived platforms, the VFS layer does not speak ARC. So data is copied from an ARC object into a VFS object, taking up space in both. If you are able to adopt your platform to use the ARC as direct VFS cache, you can save RAM that way as well.

Interesting. First that article compares GDP per capita, then explains why it is wrong and adjusts for cost of living. Then it compares those adjusted numbers to the still unadjusted numbers from europe and calls it a day.

This would be a really nice comparison, if the cost of living in any random european country happened to be exactly the median cost of living over all US states. As it stands, that article is prettified data garbage.

Yes, but in the presented hypothetical and over such a long time, the fact that it was able to maintain this in its current underdog role does not mean it would have been able to so as the main adopted 'nix.

He used launchd as his example since opinions backed by code have always had more weight in BSD communities. This led to all the systemd comparisons. But in reality, you can ignore all of that.

What his talk was about was a keynote about the ongoing paradigm shift from 'big unix iron' to ever more ephemereal OS instances that move, live, change and die faster and more rapidly. This conflicts with a number of UNIX designs where boottime matters and sets things in stone for the remainder of the current run.

These are problems that need addressing; people need to think about them and come up with ideas and solutions. The project and the community needs to decide to tackle these issues or declare them out of scope. In that context systemd is an interesting case study, since some of it tries to solve similar problems to provide a more dynamic and more MP-friendly runtime. The actually used software to accomplish this is secondary (while acknowledging that everybody has of course personal favourites and preferences).

> From what I’ve seen so far, it seems increasingly frustrating that BSD didn’t become the major free 'nix platform back in the day.

The thing is, had BSD not had the unfortunate legal incident and had been adopted as the major free 'nix platform, then it would look very different than it does today. There would have been more and different developers, different backers, different agendas. Different small influences played out over 30 years.

I love that maybe in there.

Typical being AESNI accelerated?

Just a hash or a single salted hash iteration has been on the bad practice list for a very long time. Hashes are designed to be fast, key derivation functions are intended to be slow.

There are a number of KDFs that build on top of the SHA family, from PBKDF2-HMAC-SHA1 to sha512crypt, that are all fine in principle but are often used with iteration counts that are off by 1 or 2 orders of magnitude.

5000 iterations sounds like a lot. It is not.

I understand why it is done; and that the new error pages that are not filled with technical details are a net win overall. But there are people out there that need those technical details, so that other people do not have to see them.

If you mean the one starting with D, then I may have mistyped and need to try that again. Thought that did not work for the HSTS block.

My personal grief with HSTS is that it pretty much requires you to have installed and trusted one random 400 CA collection or the other. If you have uninstalled/distrusted most of them, no current browser will allow you to access a HSTS enabled domain using one of those distrusted CAs.

There is no override to skip and accept an encryption-only connection. Which is what I would have gotten with HSTS as well, because without independent verification the CA system is a lot, but not the mutually trusted third party it claims to be.

IllumOS distros use OpenZFS as well. Unless there is a new feature flag that has not been integrated yet, FreeBSD should be able to receive from them.

rsync is therefor most likely talking about commercial Solaris with zpool versions beyond 28.

And if you end up with more than one AS due to a merge, you usually pick one them as your primary and connect the other ones to it, then announce their prefixes from your primary, ie. your primary AS becomes transit for your other ones. That way you only have to maintain one external border.

Yes, and it works good with Go since you have static executables so you do not have to set the capability on an interpreter. But you have to reapply it after every upgrade, which is easy enough to automate. That said, I mostly use mac_portacl on FreeBSD.

The bigger problem is that in go you can not do privilege revocation. And although unprivileged binding <=1023 has been a solved problem for well over a decade, most people still consider it a dark art.

So they go for a reverse proxy, more often than not, just to get the 80/443 binding.

It is things like `net/mail.ParseAddress()` that I have come to really enjoy in Go's standard library.