|
|
Subscribe / Log in / New account

Red Hat Enterprise Linux 6.2 released

Red Hat has announced the release of RHEL 6.2. "Red Hat Enterprise Linux 6.2 adds enhancements to storage and file system features including full support of iSCSI extension for RDMA. Now, benefits of low latency and high throughput through a standard SAN implementation based on 10Gb Ethernet are available to even the most demanding storage environments. This allows customers to opt out of expensive Infiniband hardware or other dedicated interconnect fabrics. Other enhancements around file system include delayed meta data logging, asynchronous and parallel file system writes, as well as support for multiple active instances of Samba in a cluster which improves overall throughput and increases availability for large Samba clustered deployments."

to post comments

Red Hat Enterprise Linux 6.2 released

Posted Dec 6, 2011 19:43 UTC (Tue) by jgg (subscriber, #55211) [Link] (37 responses)

Is it just me, or is there something wrong with the English in this press release? I've never seen something like this from an American company...

Whats the point of talking about iSER that is primarily run on IB networks in the context of getting rid of IB? Weird.

Red Hat Enterprise Linux 6.2 released

Posted Dec 6, 2011 19:47 UTC (Tue) by nix (subscriber, #2304) [Link] (36 responses)

"Other enhancements around file system include..." is a phrasing I would not expect from someone whose birth language was English unless they didn't realise that 'file system' is not a mass noun. Possibly written in .in, or written by a pure marketer with no understanding of computers?

Press releases

Posted Dec 6, 2011 19:51 UTC (Tue) by corbet (editor, #1) [Link] (34 responses)

I long since gave up on expecting any sort of coherent English out of press releases. There must be a special school somewhere that teaches the bizarre language favored for this form of writing. This particular release is not particularly bad compared to many, sadly.

Press releases

Posted Dec 6, 2011 21:53 UTC (Tue) by clugstj (subscriber, #4020) [Link] (33 responses)

Well, yes. Much market-speak in press releases. But don't they usually get the English grammar correct?

Press releases

Posted Dec 6, 2011 23:59 UTC (Tue) by ewan (guest, #5533) [Link]

Would it be very wrong to just file a bug on this?

Press releases

Posted Dec 7, 2011 0:57 UTC (Wed) by tialaramex (subscriber, #21167) [Link] (1 responses)

No, the only way a normal person will get the grammar correct for English, or really any natural language, is by being a fluent user and speaking or writing naturally.

Once they try to fit themselves to some arbitrary style guide written by people who actively despise the actual language (say, at any major newspaper, or the communications department of a major corporation) they will no longer be able to generate natural-sounding sentences, and among the unnatural results it's much harder to spot actual errors.

You might think "But, surely there are rules" and there are. But just because there _are_ rules does not mean that ordinary people using the language are conscious of what those rules are, any more than they could articulate the rules by which they walk around or catch a ball. Many smart academics have put decades of work into trying to deduce the rules by experiment, and they have made some progress. You can see a good guide to what we knew as-of the turn of the century in the book "The Cambridge Grammar of the English Language".

Press releases

Posted Dec 8, 2011 17:05 UTC (Thu) by nix (subscriber, #2304) [Link]

And what a wonderful book it is. I do wish I could get it in ebook form though (without engaging in wilful copyright violation), since it's so damn *heavy*, over two kilos, and being nearly two thousand pages is actually fairly hard to read physically.

Press releases

Posted Dec 7, 2011 11:43 UTC (Wed) by dskoll (subscriber, #1630) [Link] (29 responses)

But don't they usually get the English grammar correct?

Sadly, no. I've found that most people, including university graduates, have very poor English skills. It seems that many people just don't care. And schools and universities don't seem to care much either.

(Grumpy old man mode) When I was a kid, if you had spelling or grammatical errors in your work, you lost marks even if the subject was science or history. Nowadays, that doesn't seem to be the case; teachers accept any old garbage as long as they can vaguely make out what the student is trying to say.

Also, in universities there seems to be a culture of entitlement where students expect to be rewarded regardless of how hard they work or how well they actually do. See Pedro Pequeno's remarks from 1999.

Press releases

Posted Dec 7, 2011 13:24 UTC (Wed) by clugstj (subscriber, #4020) [Link]

Actually, my comment was limited to press releases. My experience has been that they, while quite "content neutral", are usually grammatically correct.

Press releases

Posted Dec 7, 2011 16:12 UTC (Wed) by endecotp (guest, #36428) [Link] (27 responses)

> When I was a kid, if you had spelling or grammatical errors in
> your work, you lost marks even if the subject was science or history.

That was a bad thing, and like beating children or making them work in coal mines I am glad that it has died out. If it really has died out, that is.

Press releases

Posted Dec 7, 2011 16:38 UTC (Wed) by csigler (subscriber, #1224) [Link] (1 responses)

> That was a bad thing, and like beating children or making
> them work in coal mines I am glad that it has died out.

Eye, two, am glade inforcmunt of sutch arbitraree rulez has dyed owt.

Press releases

Posted Dec 8, 2011 21:12 UTC (Thu) by jmalcolm (subscriber, #8876) [Link]

Glad to see the Scottish participating. :-)

Press releases

Posted Dec 7, 2011 16:46 UTC (Wed) by viro (subscriber, #7872) [Link]

> That was a bad thing, and like beating children or making them work in coal
> mines I am glad that it has died out. If it really has died out, that is.

Or like making the poor darlings <gasp> write legible proofs in math problems, right? Can't have that, especially when their teacher has no idea what that "proof" thing means...

And then they come to Math 101 and some poor sod has to grade their homework.
"But I wrote down the correct answer!!! How can you not give me the full marks?!?" Been there, done that...

Press releases

Posted Dec 7, 2011 17:11 UTC (Wed) by dskoll (subscriber, #1630) [Link] (23 responses)

...like beating children or making them work in coal mines...

Ummm... no. Deducting a few marks for mistakes in spelling or grammar is nothing at all like beating children or making them work in coal mines.

Furthermore, I don't think that "beating children" or "making them work in coal mines" can be glad about anything. Then again, I'd expect dangling modifiers from someone who takes your position.

Press releases

Posted Dec 8, 2011 7:19 UTC (Thu) by ekj (guest, #1524) [Link] (22 responses)

No, but it seems unfair to punish a kid with poor grades in ALL subjects for the crime of being poor in ONE.

If you're a sub-average speller, you should get subtracted points and thus potentially a lower grade in physics, math, history, biology, *everything* ?

There's a subject where students are graded for their english-skills already: it's called "english".

If the answer is not understandable, then obviously points are subtracted, but even then you don't subtract for poor english, you subtract for the fact that the student did not succeed in demonstrating knowledge of the subject area. (math, biology, chemistry or whatever)

Press releases

Posted Dec 8, 2011 9:31 UTC (Thu) by mpr22 (subscriber, #60784) [Link]

English is one of the very few cases where I'd (mostly) agree with you, since it has (IMO) the third-worst Latin-alphabet contemporary primary writing system of which I'm aware. (I put Scottish Gaelic and Irish in a tie for first place, and French a somewhat distant fourth.)

Press releases

Posted Dec 8, 2011 10:37 UTC (Thu) by dlang (guest, #313) [Link] (18 responses)

As someone who has suffered this exact problem (made worse by moving between the US and New Zealand school systems and back, with the different spelling in both), i have mixed feelings on this topic.

On the one hand, it really doesn't seem fair to be judged on spelling in technical classes. I made the exact same complaint many times while in school.

On the other hand, once you get out of school and deal with the real world, people judge you on your spelling, grammar (and capitalization/punctuation as people here reminded me a couple of weeks ago ;-) no matter what you are doing. And in many settings, if you have mistakes in these areas, people stop reading and so your entire message is lost.

Computer spelling and grammar checkers are only somewhat helpful here. They may identify problems, but they make mistakes, and so if you blindly trust them you can get even more confused messages.

So to some extent, this is a critical skill in expressing your ideas to other people, just like typing is a critical skill in getting your ideas into a computer. If you aren't good enough at it, and have to really think about it as you go, the expression of your ideas suffers. When you get "good enough" at it that you aren't thinking about the mechanics and details of it anymore, it frees up a significant chunk of brainpower that instead gets spent on the real problem.

One of the great things about working with the opensource community is that everyone seems much more willing to overlook minor glitches in this area. Part of this is that there are so many people for who english is not their primary language that mistakes are very common. But even here it is a distraction for many people.

Press releases

Posted Dec 8, 2011 10:56 UTC (Thu) by ekj (guest, #1524) [Link] (13 responses)

That's really not true.

If your job is to write text that is then published, i.e. intended to be read by a large audience, then your spelling and grammar should be good, or you'll be judged by it.

But if you hire someone for such a position - then you *should* look at their english-grades (or other demonstrations of written-communications-skill). It does not make sense to say: This person is bad in physics because he'll sometimes misspell light as ligth.

Your grade in physics, should reflect your demonstrated skill in physics. Language is a nessecary component of that, because it's what you use to demonstrate skill. But other than that, it should not matter.

Consider the case of the foreign student.

Should I get a poorer grade in *every* subject I take in a US-university, merely as a result of my english being worse than the average of natives ?

If my english is poor - give me a poor grade in *english*. Let employers who care about english-skills, judge me on that basis. (maybe that's most of them, maybe not - it depends on many things, such as where I apply for a job!)

But don't tell them: "this guy sucks in math", when that isn't true.

Language is location-dependant anyway. If I apply for a job in USA, my english would be rated as sub-average. If I apply for a job here in Norway, my english is considered substantially better than average.

Press releases

Posted Dec 8, 2011 11:04 UTC (Thu) by dlang (guest, #313) [Link] (2 responses)

Other than possibly your very first job where they _may_ ask you how you did in school. No employer cares what grades you got in school.

The only thing that matters is "did you graduate".

If you were the top of your class, you get a slight benefit, but only if it's a large, known school, so that's about 1 person in 100,000 or so. For everyone else, it doesn't matter if you graduate with a A+ average, or with a C- average, the diploma is the same.

Press releases

Posted Dec 8, 2011 11:09 UTC (Thu) by ekj (guest, #1524) [Link] (1 responses)

While that's true, it's beside the point of "should you get poor math-grades for low English-skill".

You can argue that grades generally don't matter and thus that the answer to the former question doesn't matter. That's fine, but that argument doesn't really influence the answer to the former question.

In my specific case, not only has no employer ever cared about my grades. I've not even yet ever had any employer even asking me to substantiate that I graduated at all. (when applying for a job here, you generally just state that you did, and that you'll bring evidence along for the interview - I did, but no employer ever even asked to look at it)

If you apply to continue educating yourself, grades matter. If you apply for a popular masters degree, or want to get a stipendium to peruse a ph.d, they very much *do* care about your grades.

Press releases

Posted Dec 8, 2011 12:10 UTC (Thu) by dlang (guest, #313) [Link]

if you are going after an advanced degree, you are by definition going to be publishing to a wide audience. That's part of what you have to do to get those degrees.

A PHD by definition is supposed to result in you publishing something that in a significant advancement for the field.

Press releases

Posted Dec 8, 2011 12:04 UTC (Thu) by dskoll (subscriber, #1630) [Link] (1 responses)

Consider the case of the foreign student.

I was talking about high school, not university. I agree that at the university level, teachers should take into account the student's first language and the subject matter.

When I was in university, I had a friend from Vietnam who was hired as a proofreader for a local TV magazine because his meticulousness in learning English made his English skills better than most native-born Canadians. So we can't generalize about foreign students. :)

Press releases

Posted Dec 9, 2011 14:47 UTC (Fri) by jubal (subscriber, #67202) [Link]

Ah, but that's typical for the languages you've learned to *read* first; I don't make many spelling errors, for example, but my pronunciation is frequently atrocious (and my command of the English grammar – the tenses especially – is at best shaky).

Press releases

Posted Dec 8, 2011 12:07 UTC (Thu) by dskoll (subscriber, #1630) [Link] (6 responses)

Your grade in physics, should reflect your demonstrated skill in physics. Language is a nessecary [sic] component of that, because it's what you use to demonstrate skill. But other than that, it should not matter.

I could not disagree more strongly. In just about every job I've worked at, communication skills were vital. Even in technical fields like computer programming, being able to communicate with other people is supremely important. And if you interact with customers, poor spelling and poor grammar reflect very badly on your employer.

Press releases

Posted Dec 8, 2011 12:19 UTC (Thu) by ekj (guest, #1524) [Link] (5 responses)

Communication is more than written English, it may not even happen in English at all.

Furthermore, it's fine for lots of employers to care about your English-grade. If they care about that, let them look at that.

Why force everyone, even the people who do -not- care about English-skill, to consider it anyway, because you put some fraction of it into every other grade ?

I'm not saying English-skill should not count. For a lot of jobs, especially in English-speaking countries, it obviously should count for a lot. I'm saying that your level of English-skill, should be reflected in your *English* grade.

Press releases

Posted Dec 8, 2011 12:30 UTC (Thu) by dlang (guest, #313) [Link] (2 responses)

it's not your english that's being graded in these other classes, it's your capability to communicate clearly and unambiguously.

we've all seen the humorous results of computer spelling and grammar checkers, where the computer has 'corrected' something so that the statement now means something completely different.

not correcting it means that you are playing russian roulette with your communication. It may mean what you intended it to mean, but it may not, and you don't know enough to realize this.

Press releases

Posted Dec 8, 2011 12:53 UTC (Thu) by ekj (guest, #1524) [Link] (1 responses)

I already said that:

If the answer is not understandable, then obviously points are subtracted, but even then you don't subtract for poor english, you subtract for the fact that the student did not succeed in demonstrating knowledge of the subject area.

So yes, obviously. If your English is such that it prevents you from communicating clearly and precisely about the subject-area, then you will indeed get a poorer grade, and that's perfectly fair. In that case you're not being punished for poor English as such. You're being given a lower grade because you did not succeed in demonstrating that you deserve the higher grade. (you use English for this demonstration)

I also picked examples specifically to not fall under this. If I write an answer about optics, and write about refraction of ligth instead of refraction of light, then it's perfectly clear what I mean, and I should get full credits. (not "minus a few percent")

Most minor mistakes in spelling and/or grammar fall in this category, for example in this very thread, we have one commenter who says he suffered because of differences between New Zealand and US english. Giving someone a lower grade in physics, math or history on account of "uses New Zealand spelling" is nonsense.

Press releases

Posted Dec 8, 2011 13:21 UTC (Thu) by dlang (guest, #313) [Link]

the problem is that in english, many misspellings of one word end up being a different word, which can change the meaning.

Press releases

Posted Dec 8, 2011 12:42 UTC (Thu) by anselm (subscriber, #2796) [Link]

Once you're past elementary school, your grades in English are unlikely to be based only on your mastery of English spelling, grammar and vocabulary. It turns out that the farther you advance in your education, the more English classes tend to be about things like literature, cultural history, and philosophy, and the more actual command of the English language becomes part of the »infrastructure« which is as essential to taking part in these classes as it is in math, science, or history classes.

Given this, if you're in charge of the school system and it is important to you that students have good English, it doesn't make sense to penalise people for language mistakes only in advanced English classes but not in the other subjects.

(People from countries where English is not the primary language tend to disregard this since their English classes tend to be much more focused on the language as such. Just substitute your primary language for »English« in the paragraphs above.)

Press releases

Posted Dec 8, 2011 15:38 UTC (Thu) by dskoll (subscriber, #1630) [Link]

Communication is more than written English, it may not even happen in English at all.

Feel free to substitute $YOUR_NATIVE_LANGUAGE for English in my arguments.

Why force everyone, even the people who do -not- care about English-skill, to consider it anyway, because you put some fraction of it into every other grade ?

Because communication in $YOUR_NATIVE_LANGUAGE is vital. You can be a brilliant physicist, but if you can't communicate your ideas and discoveries, your brilliance is wasted.

I'm saying that your level of English-skill, should be reflected in your *English* grade.

Yes, of course. But if your communication skills in $YOUR_NATIVE_LANGUAGE are poor, you should lose some marks in all subjects that require communication in $YOUR_NATIVE_LANGUAGE.

Press releases

Posted Dec 8, 2011 12:07 UTC (Thu) by dlang (guest, #313) [Link]

> If your job is to write text that is then published, i.e. intended to be read by a large audience, then your spelling and grammar should be good, or you'll be judged by it.

the problem is that even if you are not being formally published, you are going to be writing things intended for a reasonably wide audience, even if that audience is just your company management as you try to convince them to do something.

If you are not in a job where this matters, you are not in a job where your grades matter at all.

Press releases

Posted Dec 8, 2011 17:17 UTC (Thu) by nix (subscriber, #2304) [Link] (3 responses)

I have rather less mixed feelings on this topic than you, since I predated by only a few years the mandatory use of computers for written work production in secondary school. Of course computer use was forbidden: handwriting or nothing. If handwriting is slow and painful (I could manage 4wpm on a good day), why then you get marked down in absolutely everything and routinely castigated not just for bad handwriting but for excessive brevity and bad structure, because moving things around required writing them down again, so the first draft is the last draft. (Also a lot of teachers liked forcing us to take dictation, which meant I missed out 80% of what they were saying because I couldn't keep up, though since I couldn't read my own handwriting this was little loss).

When forced to an answer, the people imposing these rules said that good handwriting would always be essential, and that if your handwriting was not good nobody would ever read what you wrote, so it was reasonable to mark you down for every subject if your handwriting was poor. This is, you'll note, exactly the same argument as you're using here, in a slightly different domain, and it is plain that it is absolute nonsense. I never handwrite anything, nor have I in all my working life, and I have never suffered in the least for it. These days, people who handwrite are considered somewhere between eccentric and annoying, and certainly unnecessarily hard to read.

Now English spelling and grammar are harder to automate than typesetting, but it is likely that in a few years or a few decades we'll get there, and then your argument will seem as quaint and plainly flawed as my old teachers' do now.

Press releases

Posted Dec 8, 2011 17:45 UTC (Thu) by dlang (guest, #313) [Link] (2 responses)

I actually made the transition to using a computer (z80 based machine saving to cassette tapes with a dot matrix printer) in high school, and the result took my grades for papers from low C to low A

the reason was all the ease in changing things.

that being said, my poor handwriting is still a problem, when people can't read what I put on the whiteboard, or I can't make out my scribbled note from two weeks ago.

your speed in writing is like the typing speed that I see many people struggling with.

as I say, mixed feelings now, they were far from mixed at the time I was being graded.

Press releases

Posted Dec 8, 2011 19:13 UTC (Thu) by nix (subscriber, #2304) [Link] (1 responses)

I suspect we're using different definitions of 'problem'. Poor handwriting introduces constraints: you can't use whiteboards, for example. But that doesn't necessarily mean it's a problem: it's only a problem if you can't find a workaround. And it's amazing how many problems you can't fix you can find workarounds for.

But other than that I suspect we're in violent agreement.

Press releases

Posted Dec 8, 2011 19:26 UTC (Thu) by dlang (guest, #313) [Link]

Some problems you work around, but even with a work-around for a problem, it can still end up limiting you. You may decide that you don't care about that limit, you will always stay below that point, but how much of that is really that you don't (and never will) care, and how much is trying to put a good face on reality when you would otherwise feel trapped?

Lack of communication skills will definitely put a glass ceiling on your career, and you may not ever realize that it's doing so. This is the case even if you don't want to do any management type stuff.

And note that by communication skills I am talking about the ability to get your point across clearly and unambiguously. This is not talking about PC and politeness (although a lack of politeness can be an issue). If you look at many of our superstar programmers in the open source world, you will find that they are all pretty good at this communication thing, even if they are using it to send flames your way.

Press releases

Posted Dec 8, 2011 12:01 UTC (Thu) by dskoll (subscriber, #1630) [Link] (1 responses)

No, but it seems unfair to punish a kid with poor grades in ALL subjects for the crime of being poor in ONE.

Sorry, but that's life. If schools are supposed to prepare students for the real world, they'd better do it properly. If I ever came across a resume with bad spelling or poor grammar, it would greatly lower my opinion of the resume sender and make it much less likely that I'd interview him or her.

Besides, the grades subtracted for bad English were fairly low. I don't advocate failing someone in physics for substandard English, but a couple of percent off is quite appropriate.

Press releases

Posted Dec 8, 2011 17:20 UTC (Thu) by nix (subscriber, #2304) [Link]

My rough estimate of the effect of bad handwriting on my percentage-based scores in secondary school was that it halved them. A -> D, right away.

I considered this somewhat unfair at the time. I'd consider similar penalties for bad English to be somewhat unfair.

Red Hat Enterprise Linux 6.2 released

Posted Dec 7, 2011 1:53 UTC (Wed) by mbg (subscriber, #4940) [Link]

You've just hit on my bete noir: using "around" as the universal preposition (typically as a substitute for "relating to" or similar)

Believe it or not, this is a linguistic tic that is growing in popularity, at least here in Australia. It seems to have originated in public policy circles and spread as a kind of insider jargon (and don't media release writers want to be oh so cutting edge).

I suspect another reason for its virulence is that a construction using "around" allows one to riff on vague connections between things without any semantic effort to specify the nature of the relationship.

<rant off>

IPA/FreeIPA in RHEL 6.2

Posted Dec 7, 2011 6:33 UTC (Wed) by evad (subscriber, #60553) [Link] (2 responses)

In this release IPA[1] has changed from being a technology preview to a fully supported component. Has anybody had any success in deploying IPA/FreeIPA, or is now considering deploying it?

[1] http://freeipa.org/page/Main_Page

IPA/FreeIPA in RHEL 6.2

Posted Dec 7, 2011 8:17 UTC (Wed) by kashyap (guest, #55821) [Link]

There a good bunch of users/admins successfully deploying it. You might want to look at the archives of the users mailing list to follow some deployment discussions.
https://www.redhat.com/mailman/listinfo/freeipa-users

IPA/FreeIPA in RHEL 6.2

Posted Dec 11, 2011 12:02 UTC (Sun) by drag (guest, #31333) [Link]

I evaulated it this week in a VM environment. If you are willing to hand over DNS control to it and follow carefully along with the documentation and 'gotchas' then it's dead simple to setup a Kerberos/LDAP/PKI Linux domain.

How it translates to real world usage is something else. But as far as the setup it's dead simple.

Cost of 10 GbE versus IB

Posted Dec 7, 2011 11:44 UTC (Wed) by abacus (guest, #49001) [Link] (4 responses)

Since when is InfiniBand more expensive than 10 GbE DCB equipment ? A quote from blog.sandirect.com:
Then there’s Infiniband, offering four times the speed of already-speedy Ethernet, at lower cost.

Good question for historians...

Posted Dec 7, 2011 13:38 UTC (Wed) by khim (subscriber, #9252) [Link] (3 responses)

I'm not sure when it actually happened, but this article includes some numbers.

Times when Infiniband was cheaper then 10GbE Ethernet (especially the one on Cat6a cable) are year or two in the past... Today 10GbE is much cheaper and prices are still falling... I think we'll see it on desktop in the next 2-3 years...

Cost of 10 GbE versus IB

Posted Dec 7, 2011 20:15 UTC (Wed) by abacus (guest, #49001) [Link] (2 responses)

Some prices I found on the web for 10 GbE gear: and for InfiniBand equipment: As far as I can see the InfiniBand equipment has a lower price, lower latency and higher bandwidth. Or did I miss something ?

Cost of 10 GbE versus IB

Posted Dec 7, 2011 20:39 UTC (Wed) by khim (subscriber, #9252) [Link] (1 responses)

As far as I can see the InfiniBand equipment has a lower price, lower latency and higher bandwidth. Or did I miss something ?

Well, yes. 10GbE is still Ethernet - volume technology. Not only it's easier to find cheap stuff on the net (something like NC550SFP for $452.27), it's easier to get a discount (if you are buying in bulk), too.

And prices for GbE are still falling, as was already noted, for IB... not so much.

This being said right now, today, it's not always cheaper to use 10GbE as compared to IB: it all depend on the prices which you personally can manage to get. But it's only prudent to support technology which can become quite cheap later.

Note that while symmetric 10GbE switches are still rare and expensive, but asymmetric ones (a lot of 1GbE ports and couple of 10GbE ports to plug your servers into) have already arrived in bulk. This is exactly how 1GbE arrived on scene 10 years ago - and today it's dirt cheap, I don't see why 10GbE will not repeat this feat while I just can not see how IB can ever be dirt cheap: it's too specialized.

Cost of 10 GbE versus IB

Posted Dec 8, 2011 16:17 UTC (Thu) by abacus (guest, #49001) [Link]

That NC550SFP adapter doesn't support DCB, isn't it ? AFAIK DCB is necessary to avoid packet loss due to the traffic spikes caused by storage traffic.

Red Hat Enterprise Linux 6.2 released

Posted Dec 7, 2011 14:11 UTC (Wed) by dps (guest, #5725) [Link] (5 responses)

This sounds like a massive *downgrade*. If you need a different version of python then you lose because the iSCSI tools depend on a specific version of python and RHEL does not support installation of multiple versions.

Serious bastion hosts should not have any unrequired software. Using RHEL makes that very difficult. RHEL 6.2 is probably even worse than older versions.

I really can't see the use case of pure software iSCSI anyway. iSCSI HBAs do not need that support and if money is a problem then ATAoE is *much* cheaper.

Some very expensive routers support 10Gbe, and bonded 10Gbe, but not infiniband. If you have this equipment then bonded 10Gbe is cheaper than IB.

Red Hat Enterprise Linux 6.2 released

Posted Dec 7, 2011 16:26 UTC (Wed) by dsommers (subscriber, #55274) [Link]

> If you need a different version of python then you lose because the iSCSI
> tools depend on a specific version of python and RHEL does not support
> installation of multiple versions.

AFAIK, RHEL provides the complete *working* software stack, also including iSCSI. It would surprise me if this also wasn't true for 6.2. I've recently done such a setup on a 6.1 setup.

> Serious bastion hosts should not have any unrequired software. Using RHEL
> makes that very difficult. RHEL 6.2 is probably even worse than older
> versions.

RHEL is an enterprise OS. In RHEL terms, this means the supported packages will stay supported for at least 7 years since the first release major release. It basically means Red Hat is responsible for ensuring that the software packages they provide in RHEL will be supported during that whole life time.

> I really can't see the use case of pure software iSCSI anyway. iSCSI
> HBAs do not need that support and if money is a problem then ATAoE is
> *much* cheaper.

Pure software iSCSI is getting more and more used and available, even now on cheaper home NAS boxes. And Windows 7 (and probably Vista too, if somebody cares) have out-of-the-box iSCSI support too. That's something ATAoE doesn't provide so easily.

> Some very expensive routers support 10Gbe, and bonded 10Gbe, but not
> infiniband. If you have this equipment then bonded 10Gbe is cheaper than > IB.

Well, if you just measure throughput performance against a 10Gbit mark, you might be right. But if you take in consideration latencies and RDMA, infiniband is quite stronger. A single IB port may support QDR - Quad Data Rate, which is 40Gbit/s. No bonding needed.

For 10Gbe, you would need a NIC which supports RoCEE to get RDMA support. And you would need to bond 4 such NICs, to get somewhat closer to IB. But I'm not going to guarantee that this will give the exact same performance as a pure IB port in QDR, also not in latency. And I'm not sure the total price difference would be that much different, considering you probably need switches as well.

Red Hat Enterprise Linux 6.2 released

Posted Dec 7, 2011 16:39 UTC (Wed) by SEJeff (guest, #51588) [Link] (3 responses)

EPEL has a parallel installable python26 package

Red Hat Enterprise Linux 6.2 released

Posted Dec 8, 2011 10:47 UTC (Thu) by Felix (guest, #36445) [Link] (2 responses)

> EPEL has a parallel installable python26 package

That's only relevant for RHEL 5, RHEL 6 comes with Python 2.6 by default.

Though if someone (e.g. dps) needs some other version of Python, he should package that like python26 for EPEL5 instead of replacing the main Python interpreter.

Red Hat Enterprise Linux 6.2 released

Posted Dec 8, 2011 13:07 UTC (Thu) by dps (guest, #5725) [Link] (1 responses)

Duncan actively wants to minimise the software installed---the boxes in question are internet facing systems. Being forced to install things like isci-initiator tools on boxes that no use for them is very bad.

All boxes in question either have no SAN storage or full feature HBAs, so the iscsi initiator tools are pure bloat and security issues. You probably can't actually afford any of the boxen in question, although you might have talked to one.

Red Hat Enterprise Linux 6.2 released

Posted Dec 9, 2011 16:50 UTC (Fri) by raven667 (subscriber, #5198) [Link]

Wouldn't it depend on the software being SUID or running and exposed to the network to be a security issue? Just having the software installed doesn't do anything if it doesn't cross a security boundary.

You can certainly say that having extra software installed offends your sense of cleanliness but that isn't a technical issue.


Copyright © 2011, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds