Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?

Comment Re:wft ever dude! (Score 1) 211

Note that there is a difference between routing logic and forwarding logic.

The latter is arguably simplified in IPv6; the former is essentially identical.

Variable Length Addresses were demonstrated by the TUBA team in 1994, with both Cisco and Proteon demonstrating slow and fast CPU paths and hardware assistance. The cost of handling fully variable lengths was noticeable, but vanished when a common length was chosen with uncommon lengths gated, rate-limited, quenched or otherwise controlled sourcewards.

In modern forwarding engine implementations using a dual between an m-way trie and associative real memories, the cost of a full VLA is now in the noise even for arbitrary streams of random-length VLA headers; the hard part is *still* the generation of the associative arrays from the routing tries. That is, the *routing* problem is the hard problem, not the forwarding. And VLAs can simplify the routing problem if they are designed with involuntary (proxy) aggregation in mind.

The early 1990s rejection of ideas from various IPNG proposals did not anticipate a mult-decade roll-out of the minimal changes settled on in SIP+PIP (which became IPv6), nor did it have any stubs whatsoever for adjusting the on-the-wire format in the future.

This exposes the biggest single problem with the ROAD/IPNG/IPv6 process: there was almost no thought in the working groups (which became increasingly detached from operators and middle-box vendors, and were dominated by systems vendors) to deployment scenarios that were very gradual and very local, with n-level enclaves of systems with just one protocol stack (e.g., an IPv6 only bubble inside an IPv4 only bubble attached to a the Internet via an IPv6 only gateway), and the hacks that have been developed to deal with such situations (which have arisen in real life) are at least as awkward as IPv4 NAT+address overloading.

IOW, it was all end-system-software-think and little to no thinking about broader issues on end systems (ones that are multiply attached to the rest of the world, notably, or ones that migrate from one network to another rapidly), and even less about routers (especially not routers that are themselves mobile).

The slogan, "every client is also a server" should have been extended to ".. and also a first-class router", which likely would have arrived at a better overall design for IPv6, and faster deployment.

Comment Re:There is a balance between article 8 and 10 (Score 3, Informative) 401

"What kind of idiots actually write such things?"

In most of your particular extracts, it was mainly the administrators of the Marshall Plan, namely Americans and British politicians, and principally Sir David Maxwell-Fyfe MP (as he then was) taking inspiration from the work of John Peters Humphrey.

Codification was considered a good idea to avoid relitigation of common exceptions and strikings-of-balance, and to avoid imposing the need to reference foreign case law on the non-common-law countries that would agree to the document. Conflicts arose as well because some well-organized political party groupings (Christian Democrats in particular) threatened to slow ratification in a number of states simultaneously.

I think the vast majority of the people of Estonia would disagree with your assessment of the ECHR; it was a live issue in their accession referendum and is far better than the Soviet equivalent in every practical way.

Likewise, at the time it was written, Nazi laws were still on the books in the various sectors of Germany, and the legal system was a mess in all the different occupied sectors. Getting it working in *all* the sectors of occupied Germany (and Austria), including the Soviet sector was an explicit goal of the convention, and it actually succeeded in that respect for about a decade.

Finally, the document is a live one, and PACE proposes changes to clarify conflicts, to strengthen individual rights (that's the main theme) and subsidiarity, and so forth. PACE is made up of parliamentarians from each of the COE's member-states, meaning it's mostly EU parliamentarians, and since so few of their constituents engage with them on PACE, a letter written to one in an arbitrary EU member-state is likely to be looked at seriously. Maybe you could put your questions to one of them, or make some suggestions for improvement? "Just scrap it" is something they hear a lot more often from non-politicians than "fix it like this, and I'd be happier".

Comment Re:Good (Score 3, Interesting) 401

See my comment here: http://yro.slashdot.org/commen...

Roughly and in terms of English law, the ECtHR ruling upheld the Estonian Supreme Court's ruling (and that of several Estonian courts) that "L" was defamed, and that Delfi AS exacerbated the defamation by its actions, incurring a small liability for damages. Delfi admits there was the equivalent of defamation in the comments and that they were fairly treated in the Estonian court of first instance, in terms of procedure. Its argument that the Estonian law on defamation is in conflict with the ECHR has been rejected by almost everyone who has heard the case. I'd be strongly surprised if their advocates at every stage had not suggested to them that they did not have clean enough hands in the matter to pursue it through the courts with hope of success.

The ruling is not a disaster, IMHO. It tries to strike a balance for protecting the general rights of freedom of communication with the general protections from lies that are calculated to injure the reputation (and/or income and/or quiet enjoyment of life without fear), and to make striking such balances in more local courts and legislatures easier.

In brutal terms, the quantum of damages assessed by the court of first instance against Delfi was very small -- a mere slap on the wrist -- and the ECtHR took that tiny figure into account in considering the reasonabless of the law and its application. Other courts should too. Nobody went to prison, lost their business, or the like. Delfi consequently should pay costs in the appeal -- they insisted on their right to have their day in court on a small matter, and lost.

Finally, since you ask in another comment below, this would not in any way prevent someone assessed a much more severe quantum of damages (or fines, incarceration or other punitive measures) even in similar circumstances from pursuing relief through the courts, including the ECtHR. Such a person could indeed point to this case in the first instance and likely achieve a better outcome than they would have absent this decision. That's why I say it's not a disaster, not even for free speech enthusiasts. Indeed, there are some newspaper publishers in England who likely will be wishing this ruling had been made before being pressured into a deal with the late coalition government on similar matters.

Comment Re:Good thing Slashdot isn't in the EU (Score 2) 401

It's far from "unrelated" to the EU.

After the accession to the Treaty of Amsterdam in 2004 (after a popular referendum in 2003), Estonia was obliged to be a member in good standing of the European Convention on Human Rights (ECHR), the court of which (ECtHR) is only nominally independent of the European Union.

The ECtHR and ECHR are administered by the parliamentary assembly of the Council of Europe (PACE, see below), which is dominated by EU member-states and in which since the Treaty of Lisbon the EU's institutions themselves directly participate. The explicit goal of the EU is ever closer cooperation with the COE, specifically that means greater subsidiarity, as in courts across the EU that are local or specialist in nature, and legislative assemblies at all levels, will take into consideration the previous decisions of the ECtHR, and the guiding principles laid out by the PACE with respect to human rights law. The reasoning is that this will lead to less litigation and fewer appeals (which benefits individuals), and facilitate the progression to higher courts (including the ECtHR) with novel or difficult human rights questions.

This is what the Commission of the European Union says about the ECtHR:


This is what the UK Parliamentary all-party website says about the PACE. Much of the "tear up the UKHRA, withdraw from ECHR, Brexit if necessary!" faction in the UK Parliament are well aware of the work that their benchmates (and sometimes they themselves) do in PACE, which is almost always much more constructive and progressive than their public positions would suggest:


So, it's not "unrelated" because it [a] all of the member-states of the EU participate in it at all levels, [b] the EU's principal organs (the Commission, the Council, the Courts and the Parliament) and the court in question have had a formal partnership since 2012, and [c] while there are other non-EU members and observers of the COE, the COE strongly reflects the consensus of the EU and its member-states by virtue of numbers (which makes sense, as the EU itself forms the largest part of the human-geographical area with which the COE is most directly concerned).

It is only barely safe to say that the ECtHR is independent of the EU; that's still true formally, but the lines have been deliberately blurred by the EU and the COE deliberately in recent years, because that makes for more efficient administration of human rights law in the EU itself, and opens up greater access to the ECtHR by non-EU-member-states (i.e., the idea is that UK or ES cases won't clog up the ECtHR to the point where cases from UA or TR have difficulty being considered).

Comment Re: I call BS (Score 1) 184

Ah, cannot ETA, so: decent USB 3 sticks make *excellent* l2arc devices in front of spinny disk pools. They can often deliver a thousand or two IOPS each with a warm l2arc, and that would otherwise mean disk seeks. I use them regularly.

(The downside is that in almost all cases l2arc is non-persistent across reboots, crashes and pool import/export, and can take a long time to heat up so that a good chunk of seeks are being absorbed by them, so you're limited by the IOPS of the storage vdevs and the size of ARC as it heats up. That's true for *any* l2arc device, though. Building a pool out of things you would use as l2arc devices gives you persistence across import/export, reboot, and ideally panic. :-))

(Some of these downsides are likely to fade with the eventual integration of persistent l2arc in all ports, ongoing changes to arc in opensolaris, being able to assign specific DMU data to specific pool VDEVs and/or further specialization of vdevs (i.e., in addtion to log and cache), and so forth in openzfs).

Comment Re: I call BS (Score 2) 184

I've done this, experimentally, using not-super-cheap 128GiB patriot and hyper-x usb3 sticks.

For a USENET load, performance will depend on whether your incoming feed is predominantly batched, streaming, or effectively random -- small writes bother these devices individually, and aggregating them into a pool works best if you can maximize write size. One way to do that is to use a wide stripe (e.g. zpool create foo raidz stick0 stick1 stick2 stick3 stick4 ...), which works well if your load is mainly batched or streaming, such that there is lots of stuff for zfs's various write aggregation strategies to cut down on the number of small writes.

You'll also want a reasonably large ashift, for similar reasons.

If your workload is mostly reading (e.g. you have lots of downstream feeds, especially synchronous or synchronously-batched ones, or many client readers), then a sufficiently large ARC should make your choice of pool layout somewhat less important. If you are VERY read-dominant, you'll likely want to take the hit on write performance and use n-way mirroring vdevs. (e.g. zpool create foo mirror stick0 stick1 mirror stick2 stick3 ...).

If you're loss-intolerant, raidzN where N>1, or 3+-way mirroring.

What can go wrong? Mostly that you'll run out of space. :-) For that reason, a mirroring strategy, even though it's space-intensive and slows down writes, provides useful flexibility, as you can always widen or narrow mirror vdevs, or add more mirror vdevs to a pool, and it's easier to swap out vdevs in a mirror with something else than it is to swap out vdevs in a wider individual vdev (a wide raidz stripe for instance). (But you can't remove mirror vdevs (or raidzN vdevs) from a pool...).

Device names are likely be important considerations. Make sure the pool devices use device names that persist across reboots, USB3 hub failures, physical or logical disconnects and reconnects, and so forth. How to do that depends on the specific zfs port and operating system.

What can go right? Negligible seek times! That makes *all* the difference compared to spinny disks. Even for super-cheap USB 3 sticks. Really.

Final thing: no zfs port (none of the openzfs ones, and not opensolaris either) is good with USB 2 in general, and USB 2 sticks are often flaky. Avoid them.

Submission + - Hum is the world's first artificially intelligent vibrator ->

Molly McHugh writes: Hum is a smart vibrator in the ultimate sense of the word: It learns what your body likes, and it responds accordingly, delivering varying frequencies of vibrations in response to how much pressure is exerted. Unlike most vibrators, which typically only let you cycle through a few set levels of speeds and vibration patterns, Hum responds to its user’s movements to provide a unique sexual experience that mimics what it’s like to be with an actual human partner.
Link to Original Source

Comment Re:multi-culturalism (Score 1) 305

People might complain, if they knew. It's happened. Such complaints have even led to expensive litigation and ultimately informed the present UK federalism arrangement.

tl;dr it's not as simple as calling herself what she wants whenever she wants, but also nobody in any government would refuse a reasonable plan to make a change of style.

I'll restrict myself here to UK law. Laws of other countries apply to her use of styles in contexts relating to them (such as when she is physically visiting them), as well.

Formally, styles of address are the personal prerogative of the reigning monarch, and she has been delegated by Parliament the full right to style herself as she sees fit under the Royal Titles Act (1953). Exercising the prerogative _lawfully_, however, would require her to at least inform the government and have the matter Gazetted.

The statute (http://www.legislation.gov.uk/ukpga/Eliz2/1-2/9) clearly gave her the right to style herself reasonably, and that statute is probably not exhausted (there is a plausible argument that the delegated authority ceased to exist upon first use, and another that the agreement has expired because half the governments no longer share the same monarch and New Zealand introduced its own _conflicting_ statutes including the Royal Titles Act (1974, New Zealand)).

However, even if the statute is fully valid, the government-of-the-day can technically insist that she not use the delegated prerogative without formal advice from a minister. It is likely that both the department of constitutional affairs at the Ministry of Justice (which has among other things formal responsibility for policy relating to the Royal Household, the personal representatives of the monarch, and "non-delegated" royal and ecclesiastical prerogatives) and the Foreign Office (since there is still a personal union of the crowns of several independent countries) would push hard to exercise the right to give formal advice. On the other hand, most governments would simply rubber-stamp any reasonable change quickly, or even retroactively, and arrange its publication. That's important since judicial scrutiny of personal prerogatives can (and sometimes does) happen and a "slip up" by the monarch could in principle expose the government to substantial liability.

(A concrete example is in MacCormick v Lord Advocate 1953 SC 396, 1953 SLT 255 which was a case decided under Scottish Law involving the current monarch's choice to style herself Elizabeth II. The courts found, essentially, that the monarch had acted lawfully in styling herself Elizabeth II, and the decision refers to minutes and letters by government ministers endorsing her name(s), that the supporting statute was valid with respect to MacCormick & Hamilton's claim, and that her regnal name was widely published, and in particular was Gazetted here https://www.thegazette.co.uk/L... -- note that "Published by Authority" means by the authority of the government, not the monarch, and "with the advice of our privy council" means that there was a formal consultation with the government.

MacCormick was a Scottish Nationalist lawyer and Rector of the University of Glasgow, and raised the matter as a point of interesting differences in Scottish and English constitutional and administrative law; Hamilton is a well known Scottish Nationalist lawyer; both were trying to establish that Scotland had the right to be consulted on matters of personal prerogative. In effect, although they lost this particular case, their ideas are now entrenched in devolution law -- a change to the monarch's style would almost certainly have to be supported by an Act of the Scottish Parliament. In particular, MacCormick was specifically referenced in the parliamentary debates leading to paragraphs (1) and (2) of Part I of Schedule 5 of the Scotland Act (1988) -- it is likely that in litigation in the Supreme Court of the United Kingdom, the monarch's style (in Scotland) would be found to be an unreserved matter under (2)(a) rather than covered by (1)(a) or implied in (3)(3)(a).)

So in short, Unelected Ceremonial President Windsor has to abide by various Settlements which require her to give up many of her personal freedoms (speech and expression, religion, association, full free choice in marriage) in exchange for continuing use of vast tracts of land, palaces, money, jewels, publicly funded servants, and so forth. That's been the case in most constitutional and even many absolute monarchies in Western Europe since no later than the 15th century.

Comment Re:The US needs a loser-pays legal system (Score 4, Informative) 136

The current system (in U.S. District courts) *is* loser pays (see U.S. F.R.Civ.P. rule 54(d)).

Where the federal courts differ from most "loser pays" systems is that evidence of offers to settle ahead of a trial is generally excluded as a matter of policy.

Pretty much no loser-pays system (and that includes federal courts and several private law systems in the various states) actually requires that the loser *always* pays the full costs of the other side *in all circumstances*; wide latitudes are given to the courts to assess costs in a way it feels is just, or appropriate to the behaviour of the parties, etc. U.S. district courts have narrower latitude than both, owing in part to statute.

Generally speaking, if no offers to settle out of court are made (and thus also not rejected), then the loser generally is assessed costs unless it would be unjust to do so, thus "loser pays". However, offers to settle out of court are normal and even in district courts are encouraged to avoid unnecessary court costs and time dealing with controversies which can be worked out by the litigating parties outside of court.

In most loser-pays systems costs are assessed against parties who should have ended litigation sooner. For example under most systems that use a regulated offer along the lines of the Calderbank rules (this is definitely untrue of many state systems and U.S. district courts, but is true in some states, such as Florida), a winning party that was made an offer to settle out of court that it rejected and subsequently did not beat in court is usually assessed at least some proportion of the offeror's costs after that point, even if the offeror is ultimately the losing party. That is, even though the party won, it could have achieved the same result with fewer costs to the parties and the courts, and should therefore bear some of those avoidable costs. There are often codified forms of offer which make it even more clear that refusing a well-pitched offer could be expensive (as in Part 36 of the Civil Procedure Rules (England and Wales)) for a party that does not subsequently better it.

Additionally, most systems allow the parties to agree on how to split costs in order to avoid further litigation on who should pay which costs; the motions under F.R.Civ.P rule 54(d)(1)&(2) are frequently consent motions agreed between the parties after judgement.


Comment Re:interesting... (Score 1) 417

Here's a nice bit of history about religious tolerance and liberty in the colonies right in the middle of the American Revolution:


"It is difficult to overestimate the degree to which, on the eve of the Revolution, Catholics in America were still widely discriminated against. Several members of the Continental Congress, including Congregationalist Roger Sherman, were opposed to hiring Catholics to fight in the Continental Army. Only three colonies allowed Catholics to vote. They were banned from holding public office in all New England colonies save Rhode Island. New Hampshire law called for the imprisonment of all persons who refused to repudiate the pope, the mass, and transsubstantiation. New York held the DEATH PENALTY [emphasis mine] over priests who entered the colony; Virginia boasted that it would only arrest them."

In Virginia, the birthplace of the separation of church and state, it took *seven years* for Thomas Jefferson to convince the General Assembly to pass the Virginia Statute for Religious Freedom, and debates on the matter bear a striking resemblance to the sorts of thing one might read in YouTube comments.

By the time that the United States Bill of Rights was ratified, the freedom to practice any religion without fear of being barred from holding land, accessing the courts, or holding most professional jobs had been established by law in most of the British Empire.

This is not entirely surprising as many of the most influential people who formed the Federalist faction in what became the United States were in close cooperation with the Foxites in the British parliament from well before the Revolution until well after, and agreed on many -- or even most -- civil liberties and constitutional issues. The American Revolution weakened the common enemy (principally the Northites and Grenvilleites, who are all fairly called Tories in spite of their claim to the Whig mantle).

By comparison, the erosion of Tory (see above) dominance in the British parliament in the wake of the Seven Years' War led to a series of religious Relief Acts relaxing restrictions on Catholics. It's noteworthy that the first major such act, the Quebec Act 1774, was one of the "Intolerable Acts" protested by the Americans (in the political faction sense) that they argued justifed Independence. Additionally, in the thick of the Revolution, the British parliament passed the Relief Act 1778 and the Schools and Bishops Act 1782, in spite of vigorous domestic opposition (there were riots in Britain in the wake of each), and even more vigorous opposition in the parts of the Thirteen Colonies not already in full rebellion, and some upset in several of the others that ultimately did not join the American Revolution.

Comment Re:if you want a trusted proxy.. (Score 1) 177

Also, by way of self-followup, the internet-draft proposes a mechanism to distinguish between https URIs over encrypted http2 connections and http URIs over encrypted http2 connections, with the goal that only the latter will be subject to manipulation by an explicitly trusted proxy, and that the client, the server, and conforming proxies all take steps to avoid unnoticed manipulation or examination of https URI related data. (The client, server and proxy all have opportunities to "opt out" of the proxy's manipulation or examination of http URI related content).

Comment Re:if you want a trusted proxy.. (Score 1) 177

Sure, Terry, but what's worse, an MITM DOS ("you don't get to negotiate a [n https] connection") or an MITM that allows full inspection and modification of data that one (typically the server-side) or both ends think is an HTTPS connection?

The server side has lots of standardized and/or developed tools to protect the integrity and privacy of data between the server and the browser that does not rely upon perfect (or even any) HTTPS. The assumption that HTTPS is perfect -- or even close enough -- has been holding back deployments of such tools for more than a decade.

Apart from opening doors to greater network efficiency along several axes (caching and other deduplication and localization approaches, distribution away from single front ends, greater concentration of resources on single NLA addresses), deliberate trusted proxy ought to push people into reconsidering whether only-the-server-side-has-a-certificate TLS is *sufficient* for integrity and privacy. (I don't think it is, and I suspect that view is shared by at least some people behind two-factor authentication etc.)

However, the safer bet is that the proposal is likely to be strangled by people with a very narrow view of HTTPS, or by a lack of engagement on the way deliberate trusted MITM affects the security model of the whole WWW (few people will ask if it can actually *improve* overall security compared to the many people who will argue that it necessarily erodes it). It's a pity, because separating integrity (arbitrary chunks of data with some signed checksum) and privacy is likely to be a clear win on energy where integrity-but-not-privacy-required data are popular enough to warrant being highly distributed and/or highly cached.

Finally, another question that ought to be addressed is that a lot of eggs are in the HTTPS basket, and not all of them have been inherited from everything-over-HTTP. That deliberate trusted MITM exposes this question is not a bad thing, I think. However I would again bet that the consensus will be that the convenience of everything-over-HTTPS will trump that of a standard approach to making (especially) HTTP inside HTTPS amenable to caching, ad-removal, virus-removal, and whatever else a "middlebox" might do with HTTP now. After all, if someone really wants to do that, they can simply disallow (well, "break", even) HTTPS negotiation altogether...

Thus, my own answer to my question at the top: where there is agreement to enable a trusted MITM to act, then an MITM DOS is much worse; but for any other case it's the lesser of two evils. The key thing here is in how to determine agreement and trust; where that's not clear to either the client side or the server side (which is likely a harder problem), surely it's not enormously different from an MITM DOS ("can't establish a valid HTTPS connection" because of anything from TCP port blocking to pinned certificate mismatching).

Comment Re:Freakin' Riders. (Score 1) 767

They aren't riders in the U.S. sense in any meaningful way because of sec. 54 of the Constitution Act 1867, which gives the government -- even a minority government -- the exclusive right to introduce (into the House of Commons per sec. 53) money bills:

It shall not be lawful for the House of Commons to adopt or pass any Vote, Resolution, Address, or Bill for the Appropriation of any Part of the Public Revenue, or of any Tax or Impost, to any Purpose that has not been first recommended to that House by Message of the Governor General in the Session in which such Vote, Resolution, Address, or Bill is proposed

That is, *first* recommended.

The rules of both houses of Parliament impose further restrictions on their members with respect to amendments.

Omnibus bills are a problem, certainly, but it is one fully under the control of successive governments and tolerated by the House of Commons (who would force an election in rejecting or heavily amending a money bill), even during the recent series of minorities and weak majorities.

The federal government further enjoys several powers to veto legislation that one house or even all of Parliament passes anyway, even when they control only a minority of seats in the House of Commons.

Omnibus bills are far from new, and the Canadian system was sufficiently weak that the Australian constitution (which granted similar exclusive rights over the introduction of money bills and the disposition of all bills) added section 55 to their equivalent Act:

Laws imposing taxation shall deal only with the imposition of taxation, and any provision therein dealing with any other matter shall be of no effect.

                                      Laws imposing taxation, except laws imposing duties of customs or of excise, shall deal with one subject of taxation only; but laws imposing duties of customs shall deal with duties of customs only, and laws imposing duties of excise shall deal with duties of excise only.

Similar proposals have been debated in Canada for more than a century, however since Omnibus budget bills are under the control of the federal government, there has never been much headway made, and governments have tended to consolidate and split bills according to their own needs and pushback from the Parliamentary committees and agencies who review the bills. "Omninbussing" follows trends, and somewhat reflect the strength of party discipline across the whole of Parliament -- a relatively weak government is often the source of large and complicated money bills, while a government controlling both Houses and enjoying strong party discipline tends to produce more and slimmer bills.

Riders are a wholly different matter.

The U.S. system permits riders since the Executive can only exercise a Presidential veto. The only additional checks on riders -- arbitrary amendments made by majority vote in *either* house of Congress -- are the Origination Clause, bicameralism generally, and the rules of the House of Representative, which all serve to limit the damage any subset of Congress, particularly Senators, can impose on a budget bill. In practice, however, all sides compromise and allow arbitrary changes to be added by various factions on critical bills that the Executive cannot afford to veto, so the system is more a "balance" than a "check" on legislators. Successive House majorities have over *centuries* softened their stance on the exclusivity of their control over the budget process, by weakening their own interpretation of the Origination Clause, by compromising with Senators and House factions, and by continuing to present to the President bills that are difficult to veto. On the other hand, the Executive has adapted by lobbying individual members of Congress directly and aggressively, and various processes have evolved such that the Executive can effectively derail items in money bills that it objects to; most of those exploit the essentially permanent weak party discipline in Congress.

Finally, you wrote, "99% of the time, the regulations and laws enacted through such dirty tactics are things which would never be approved if they were voted on as single items". However, the standing committees in both the House of Commons and the Senate do in fact typically vote on each individual item; contentious matters are further voted on individually in Committees of the Whole rather than splitting the matters out into individual bills (which would require new Royal Recommendations). However, all these votes are won by a majority, and when a government's supporters dominate all of Parliament, they are practically foregone conclusions, as individual internal party factions' disagreements are typically dealt with in advance by MPs and Senators lobbying ministers directly or in caucus meetings.

Things that reasonably could be called "dirty" or at least highly partisan and essentially severable matters did find their way into money bills during several recent minority governments, and survived passing through a fractious Houses of Commons and even a Senate controlled by opponents. All parties agreed to compromise rather than to force an electoral contest over each (or any) such matter, although most of them passed essentially unchanged. Indeed, the motion of no confidence that triggered the 2006 federal election was not triggered by the a conflict over a particular piece of legislation, in spite of collapsed deals on compromises between the government and the opposition parties.

Finally, on the matter of riders and budgets, the Clark government collapsed in 1979 because he had not pleased enough of the various factions of the House of Commons. The formal mechanism was an amendment to a money bill that was insisted upon by a majority of the MPs present, and that amendment was simply a wrecker ("... that this House has lost confidence in the government") in part to avoid running into the Royal Recommendation rule that additionally prevented a potentially saving amendment proposed by a minor opposition party which would have required one in advance. The minor opposition party was, amusingly, obliterated in the 1980 election when the créditistes lost all five of their seats to Liberals. Had the Canadian federal system permitted riders in the U.S. sense, and the sense I believe you mean, the Clark government likely would have continued for perhaps another year, and would have faced a Liberal party led by someone other than Trudeau.

Comment Re: Advatages of ZFS over BTRFS? (Score 1) 297

L2ARC is slowly populated from the older ends of the ARC, using a separate thread. Not all blocks aging out of the ARC are guaranteed to get to the L2ARC. This is to avoid slowdowns when the ARC is under pressure, and to allow for the use of cache vdevs on media which are slow to write but fast to read.

L2ARC vdevs are circular buffers.

Each L2ARC entry consumes at least 50 bytes from the ARC, and may consume considerably more. Those bytes are released ONLY after the circular buffer overwrites the L2ARC entry. Therefore a large L2ARC competes for ARC space with ordinary blocks, and the larger the L2ARC the more likely it is that the L2ARC contains stale blocks. Those stale blocks continue to use ARC space.

The L2ARC's big use is to avoid seeks to fetch occasionally-used data. It helps so little with streaming reads that data that was streamed (via prefetch mechanisms, for example) into the ARC are skipped by the thread that populates L2ARC vdevs. Defeating that (forcing the storage of streamed data) typically worsens performance of a pool.

The ZIL is automatically managed areas in storage vdevs in each pool. One can configure one or more separated log vdevs which will be used instead of the ZIL. When synchronous write calls are made, the data is stored in the slog (if available) or ZIL, and the call returns. No other data is written to the slog or ZIL. The blocks written out remain in the ARC, marked as dirty and in need of an asynchronous flush. When the open transaction group closes, the in-ARC copies of what went to the slog or ZIL is written out with everything else in the txg, and then the slog or ZIL is cleared.

Neither the slog nor the ZIL is read except during the import process; if they are non-empty at import, the blocks are written out (synchronously) to the pool's storage vdevs and then cleared.

"With the ZIL in a different drive (SSD or otherwise), you reduce the number of writes required"

The ZIL is the ZIL; the slog is the slog. If you have a slog you don't write to the ZIL, and whether that changes the number of writes of *synchronous* data is configuration-dependent. If you have mirrored slogs, for example, you are probably writing more than you would if you just used the ZIL.

In either case the idea is to write out synchonous blocks quickly and with as little writing latency as possible; writes are linear and are to areas at the start of the vdevs.

"Because you can generally write to a ZFS pool significantly faster than to a single disk"

Actually, it's the other way around, but hinges on what you mean by "write". The more disks there are in the pool the more labels have to be updated at the finalization of each txg. That's not a large extra amount to push out to the rotating material, but it's done synchronously and will *invariably* result in seeks to the start and end of each component device in each vdev in the pool.

Additionally, you can configure a single-disk pool, and there are reasons why you might want to do so, even though that is UNSAFE.

However, system calls return quickly because all writes go into the ARC, and asynchronous write calls can return immediately; synchronous write calls return when the data is committed to the ZIL or slog(s). In the case where there's a slog, write calls practically never initiate actual activity on the device(s) forming the storage vdev(s). Instead, writes will be triggered by timers.

For many workloads, this makes writes to pools seem very fast, since delayed writing allows for smarter scheduling, as well as coalescing of writes to specific physical blocks.

Don't sweat it -- it's only ones and zeros. -- P. Skelly