Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.debian.user > #212344 > unrolled thread
| Started by | Steven Mainor <steve@degga.net> |
|---|---|
| First post | 2019-08-07 07:40 +0200 |
| Last post | 2019-08-09 12:20 +0200 |
| Articles | 20 on this page of 84 — 32 participants |
Back to article view | Back to linux.debian.user
Server hardware advice. Steven Mainor <steve@degga.net> - 2019-08-07 07:40 +0200
Re: Server hardware advice. Richard Hector <richard@walnut.gen.nz> - 2019-08-07 08:00 +0200
Re: Server hardware advice. Steven Mainor <steve@degga.net> - 2019-08-07 08:10 +0200
Re: Server hardware advice. Nektarios Katakis <nektarios@mail.nektarioskatakis.xyz> - 2019-08-07 12:20 +0200
Re: Server hardware advice. mick crane <mick.crane@gmail.com> - 2019-08-07 14:10 +0200
Re: Server hardware advice. deloptes <deloptes@gmail.com> - 2019-08-07 17:00 +0200
Re: Server hardware advice. Michael Stone <mstone@debian.org> - 2019-08-07 17:10 +0200
Re: Server hardware advice. deloptes <deloptes@gmail.com> - 2019-08-07 17:20 +0200
Re: Server hardware advice. Michael Stone <mstone@debian.org> - 2019-08-07 17:40 +0200
Re: Server hardware advice. ghe <ghe@slsware.net> - 2019-08-07 17:50 +0200
Re: Server hardware advice. Celejar <celejar@gmail.com> - 2019-08-08 00:20 +0200
Re: Server hardware advice. Steven Mainor <steve@degga.net> - 2019-08-07 18:50 +0200
Re: Server hardware advice. deloptes <deloptes@gmail.com> - 2019-08-08 07:50 +0200
Helpful attitude (was: Server hardware advice.) Nicolas George <george@nsup.org> - 2019-08-08 11:00 +0200
Re: Helpful attitude (was: Server hardware advice.) Kenneth Parker <sea7kenp@gmail.com> - 2019-08-08 12:50 +0200
Re: Helpful attitude (was: Server hardware advice.) Jonas Smedegaard <jonas@jones.dk> - 2019-08-08 14:10 +0200
Re: Helpful attitude (was: Server hardware advice.) Dan Ritter <dsr@randomstring.org> - 2019-08-08 16:20 +0200
Re: Helpful attitude (was: Server hardware advice.) Jonas Smedegaard <jonas@jones.dk> - 2019-08-08 17:00 +0200
Re: Helpful attitude (was: Server hardware advice.) deloptes <deloptes@gmail.com> - 2019-08-08 17:20 +0200
Re: Helpful attitude (was: Server hardware advice.) Jonas Smedegaard <jonas@jones.dk> - 2019-08-08 18:00 +0200
Re: Helpful attitude (was: Server hardware advice.) Judah Richardson <judahrichardson@gmail.com> - 2019-08-08 21:50 +0200
Re: Helpful attitude (was: Server hardware advice.) deloptes <deloptes@gmail.com> - 2019-08-09 00:50 +0200
[OT] Helpful attitude (was: Server hardware advice.) <tomas@tuxteam.de> - 2019-08-09 10:00 +0200
Re: [OT] Helpful attitude (was: Server hardware advice.) Étienne Mollier <etienne.mollier@mailoo.org> - 2019-08-09 11:20 +0200
Re: [OT] Helpful attitude (was: Server hardware advice.) <tomas@tuxteam.de> - 2019-08-09 11:50 +0200
Re: [OT] Helpful attitude (was: Server hardware advice.) Étienne Mollier <etienne.mollier@mailoo.org> - 2019-08-09 12:30 +0200
Re: [OT] Helpful attitude (was: Server hardware advice.) tomas@tuxteam.de - 2019-08-09 14:40 +0200
Re: [OT] Helpful attitude (was: Server hardware advice.) Étienne Mollier <etienne.mollier@mailoo.org> - 2019-08-09 15:50 +0200
Re: [OT] Helpful attitude John Hasler <jhasler@newsguy.com> - 2019-08-09 15:50 +0200
Re: [OT] Helpful attitude The Wanderer <wanderer@fastmail.fm> - 2019-08-09 16:00 +0200
Re: [OT] Helpful attitude Larry Martell <larry.martell@gmail.com> - 2019-08-09 16:10 +0200
Re: [OT] Helpful attitude Étienne Mollier <etienne.mollier@mailoo.org> - 2019-08-09 16:10 +0200
Re: [OT] Helpful attitude John Hasler <jhasler@newsguy.com> - 2019-08-09 17:30 +0200
Re: [OT] Helpful attitude Étienne Mollier <etienne.mollier@mailoo.org> - 2019-08-09 19:00 +0200
Re: [OT] Helpful attitude deloptes <deloptes@gmail.com> - 2019-08-09 18:30 +0200
Re: [OT] Helpful attitude Brian <ad44@cityscape.co.uk> - 2019-08-09 20:50 +0200
Re: Helpful attitude (was: Server hardware advice.) Curt <curty@free.fr> - 2019-08-08 22:20 +0200
Re: Helpful attitude Richard Owlett <rowlett@cloud85.net> - 2019-08-09 16:10 +0200
Re: Helpful attitude (was: Server hardware advice.) ghe <ghe@slsware.net> - 2019-08-09 08:20 +0200
Re: Helpful attitude (was: Server hardware advice.) Gene Heskett <gheskett@shentel.net> - 2019-08-09 09:20 +0200
Re: Helpful attitude (was: Server hardware advice.) "Bob Crochelt" <rfc56@fastmail.com> - 2019-08-09 14:00 +0200
Re: Helpful attitude (was: Server hardware advice.) Gene Heskett <gheskett@shentel.net> - 2019-08-09 16:20 +0200
Re: Helpful attitude (was: Server hardware advice.) "Bob Crochelt" <rfc56@fastmail.com> - 2019-08-10 00:00 +0200
Re: Helpful attitude (was: Server hardware advice.) rhkramer@gmail.com - 2019-08-08 14:10 +0200
Re: Server hardware advice. Reco <recoverym4n@enotuniq.net> - 2019-08-07 09:00 +0200
Re: Server hardware advice. john doe <johndoe65534@mail.com> - 2019-08-07 09:40 +0200
Re: Server hardware advice. Jonas Smedegaard <jonas@jones.dk> - 2019-08-07 10:50 +0200
Re: Server hardware advice. Jonas Smedegaard <jonas@jones.dk> - 2019-08-07 10:30 +0200
Re: Server hardware advice. Reco <recoverym4n@enotuniq.net> - 2019-08-07 11:00 +0200
Re: Server hardware advice. Jonas Smedegaard <jonas@jones.dk> - 2019-08-07 11:10 +0200
Re: Server hardware advice. Steven Mainor <steve@degga.net> - 2019-08-07 12:10 +0200
Re: Server hardware advice. Jonas Smedegaard <jonas@jones.dk> - 2019-08-07 12:50 +0200
Re: Server hardware advice. jochen-2019-q2@ozark.de - 2019-08-07 22:40 +0200
Re: Server hardware advice. Andrei POPESCU <andreimpopescu@gmail.com> - 2019-08-08 09:30 +0200
Re: Server hardware advice. John Hasler <jhasler@newsguy.com> - 2019-08-09 05:20 +0200
Re: Server hardware advice. Steven Mainor <steve@degga.net> - 2019-08-09 08:00 +0200
Re: Server hardware advice. John Hasler <jhasler@newsguy.com> - 2019-08-09 15:40 +0200
Re: Server hardware advice. deloptes <deloptes@gmail.com> - 2019-08-09 18:20 +0200
Re: Server hardware advice. Reco <recoverym4n@enotuniq.net> - 2019-08-09 19:00 +0200
Re: Server hardware advice. Celejar <celejar@gmail.com> - 2019-08-09 19:20 +0200
Re: Server hardware advice. Reco <recoverym4n@enotuniq.net> - 2019-08-09 20:30 +0200
OT "x times cheaper", was: Re: Server hardware advice. Richard Hector <richard@walnut.gen.nz> - 2019-08-10 10:00 +0200
Re: OT "x times cheaper", was: Re: Server hardware advice. deloptes <deloptes@gmail.com> - 2019-08-10 11:20 +0200
Re: OT "x times cheaper", was: Re: Server hardware advice. Richard Hector <richard@walnut.gen.nz> - 2019-08-10 11:30 +0200
Re: OT "x times cheaper", was: Re: Server hardware advice. David Wright <deblis@lionunicorn.co.uk> - 2019-08-10 17:10 +0200
Re: OT "x times cheaper", was: Re: Server hardware advice. Richard Hector <richard@walnut.gen.nz> - 2019-08-11 00:50 +0200
Re: OT "x times cheaper", was: Re: Server hardware advice. mick crane <mick.crane@gmail.com> - 2019-08-11 09:50 +0200
Re: OT "x times cheaper", was: Re: Server hardware advice. Curt <curty@free.fr> - 2019-08-11 12:00 +0200
Re: Server hardware advice. John Hasler <jhasler@newsguy.com> - 2019-08-09 19:50 +0200
Re: Server hardware advice. Steven Mainor <steve@degga.net> - 2019-08-10 04:00 +0200
Re: Server hardware advice. Steven Mainor <steve@degga.net> - 2019-08-07 12:00 +0200
Re: Server hardware advice. Reco <recoverym4n@enotuniq.net> - 2019-08-07 13:00 +0200
Re: Server hardware advice. David Christensen <dpchrist@holgerdanske.com> - 2019-08-08 09:10 +0200
Re: Server hardware advice. John Hasler <jhasler@newsguy.com> - 2019-08-08 15:20 +0200
Re: Server hardware advice. Celejar <celejar@gmail.com> - 2019-08-08 15:50 +0200
Re: Server hardware advice. Dan Ritter <dsr@randomstring.org> - 2019-08-08 16:30 +0200
Re: Server hardware advice. Jonas Smedegaard <jonas@jones.dk> - 2019-08-08 17:00 +0200
Re: Server hardware advice. Reco <recoverym4n@enotuniq.net> - 2019-08-08 17:30 +0200
Re: Server hardware advice. Jonas Smedegaard <jonas@jones.dk> - 2019-08-08 18:20 +0200
Re: Server hardware advice. Steven Mainor <steve@degga.net> - 2019-08-08 22:50 +0200
Re: Server hardware advice. Dan Ritter <dsr@randomstring.org> - 2019-08-08 23:10 +0200
Re: Server hardware advice. Michael Stone <mstone@debian.org> - 2019-08-08 23:20 +0200
Re: Server hardware advice. David Christensen <dpchrist@holgerdanske.com> - 2019-08-09 00:10 +0200
Re: Server hardware advice. Igor Cicimov <icicimov@gmail.com> - 2019-08-09 12:20 +0200
Page 4 of 5 — ← Prev page 1 2 3 [4] 5 Next page →
| From | Reco <recoverym4n@enotuniq.net> |
|---|---|
| Date | 2019-08-09 20:30 +0200 |
| Message-ID | <ythOi-607-3@gated-at.bofh.it> |
| In reply to | #212524 |
On Fri, Aug 09, 2019 at 01:16:49PM -0400, Celejar wrote: > On Fri, 9 Aug 2019 19:59:34 +0300 > Reco <recoverym4n@enotuniq.net> wrote: > > > On Fri, Aug 09, 2019 at 06:16:21PM +0200, deloptes wrote: > > ... > > > > This one was very appealing > > > https://www.amazon.de/DMC-Taiwan-Industrial-Networking-Processor/dp/B07T3TWYLJ/ref=sr_1_11?__mk_de_DE=%C3%85M%C3%85%C5%BD%C3%95%C3%91&keywords=acrosser&qid=1561237800&s=gateway&sr=8-11 > > > > $430 for a router? Surely you're kidding. > > > > This one is five times cheaper *and* it can run Debian or openwrt: > > > > https://www.amazon.com/Linksys-Dual-Band-Wireless-Gigabit-WRT1200AC/dp/B00UVN20T0/ref=sr_1_2?keywords=linksys+wrt+1200&qid=1565369861&s=gateway&sr=8-2 > > When you say five times cheaper, I gather you're talking about the > prices for used units, in which case it's not really an > apples-to-apples comparison. At least when I checked, the new units on > Amazon start at $190. I stand corrected. It's two times cheaper for new ones. Reco
[toc] | [prev] | [next] | [standalone]
| From | Richard Hector <richard@walnut.gen.nz> |
|---|---|
| Date | 2019-08-10 10:00 +0200 |
| Subject | OT "x times cheaper", was: Re: Server hardware advice. |
| Message-ID | <ytus9-5wf-5@gated-at.bofh.it> |
| In reply to | #212529 |
[Multipart message — attachments visible in raw view] — view raw
On 10/08/19 6:20 AM, Reco wrote: > On Fri, Aug 09, 2019 at 01:16:49PM -0400, Celejar wrote: >> When you say five times cheaper, I gather you're talking about the >> prices for used units, in which case it's not really an >> apples-to-apples comparison. At least when I checked, the new units on >> Amazon start at $190. > > I stand corrected. It's two times cheaper for new ones. <rant> Sorry, this usage grates with me. $amount cheaper that $price means subtract $amount from $price $x times $price means multiply $price by $x so "2 times cheaper (than $450)" is: $450 - (2 x $450) = -$450. </rant>
[toc] | [prev] | [next] | [standalone]
| From | deloptes <deloptes@gmail.com> |
|---|---|
| Date | 2019-08-10 11:20 +0200 |
| Subject | Re: OT "x times cheaper", was: Re: Server hardware advice. |
| Message-ID | <ytvHz-6te-13@gated-at.bofh.it> |
| In reply to | #212560 |
Richard Hector wrote: > <rant> > Sorry, this usage grates with me. > > $amount cheaper that $price means subtract $amount from $price > > $x times $price means multiply $price by $x > > so "2 times cheaper (than $450)" is: > > $450 - (2 x $450) = -$450. so what multiplied by 2 gives 450? 450 is 100% or 1 225 is 50% or 1/2 perhaps this is the confusion, cause we are using daily language to refer to maths. In fact I would do it the other way around. initial price x 1xtime x+(1*x) 2xtimes x+(2*x) this gives x=150 The problem is in the comparison - when you read the characteristics of the devices, there are couple of differences that can not be neglected, but I agree 400 is too much. I think fair price for such a device would be in 200-300 range, what we also see for the Lynksis router.
[toc] | [prev] | [next] | [standalone]
| From | Richard Hector <richard@walnut.gen.nz> |
|---|---|
| Date | 2019-08-10 11:30 +0200 |
| Subject | Re: OT "x times cheaper", was: Re: Server hardware advice. |
| Message-ID | <ytvRf-6wE-7@gated-at.bofh.it> |
| In reply to | #212568 |
[Multipart message — attachments visible in raw view] — view raw
On 10/08/19 9:10 PM, deloptes wrote: > Richard Hector wrote: > >> <rant> >> Sorry, this usage grates with me. >> >> $amount cheaper that $price means subtract $amount from $price >> >> $x times $price means multiply $price by $x >> >> so "2 times cheaper (than $450)" is: >> >> $450 - (2 x $450) = -$450. > > so what multiplied by 2 gives 450? > > 450 is 100% or 1 > 225 is 50% or 1/2 Right, so 225 is 50% cheaper, or half cheaper. Not twice cheaper. > perhaps this is the confusion, cause we are using daily language to refer to > maths. Daily language is the problem, yes. I'm not saying my fight is an easy one :-) > In fact I would do it the other way around. > > initial price x > 1xtime x+(1*x) > 2xtimes x+(2*x) > > this gives x=150 450 is two times more expensive than 150 (or 200% more than), or three times as expensive as 150 (or 300% as expensive). 300 is two times as expensive as 150, or 100% more expensive than 150 We know that these don't work symmetrically; if you have a 50% discount, you can't get the original price back by adding 50%, because it's 50% of a different number. Richard
[toc] | [prev] | [next] | [standalone]
| From | David Wright <deblis@lionunicorn.co.uk> |
|---|---|
| Date | 2019-08-10 17:10 +0200 |
| Subject | Re: OT "x times cheaper", was: Re: Server hardware advice. |
| Message-ID | <ytBai-1y7-7@gated-at.bofh.it> |
| In reply to | #212569 |
On Sat 10 Aug 2019 at 21:19:31 (+1200), Richard Hector wrote: > On 10/08/19 9:10 PM, deloptes wrote: > > Richard Hector wrote: > > > >> <rant> > >> Sorry, this usage grates with me. > >> > >> $amount cheaper that $price means subtract $amount from $price > >> > >> $x times $price means multiply $price by $x > >> > >> so "2 times cheaper (than $450)" is: > >> > >> $450 - (2 x $450) = -$450. > > > > so what multiplied by 2 gives 450? > > > > 450 is 100% or 1 > > 225 is 50% or 1/2 > > Right, so 225 is 50% cheaper, or half cheaper. Not twice cheaper. > > > perhaps this is the confusion, cause we are using daily language to refer to > > maths. > > Daily language is the problem, yes. I'm not saying my fight is an easy > one :-) > > > In fact I would do it the other way around. > > > > initial price x > > 1xtime x+(1*x) > > 2xtimes x+(2*x) > > > > this gives x=150 > > 450 is two times more expensive than 150 (or 200% more than), or three > times as expensive as 150 (or 300% as expensive). > 300 is two times as expensive as 150, or 100% more expensive than 150 > > We know that these don't work symmetrically; if you have a 50% discount, > you can't get the original price back by adding 50%, because it's 50% of > a different number. "Expensive" is a dimensional term, like length and time. "Cheap" is in a different category, like shortness. A 6-inch nail is twice as long as a 3-inch nail, but one doesn't say the latter is twice as short. But if someone asked for a nail twice as short as this (holding up a 6-inch nail), you might assume they were a non-native speaker of English, or you might notice you're almost twice as tall as they are: ie it's a child. (And it would be polite to offer them a 3-inch nail. Learning all the categories takes time, and some people might have slightly different boundaries.) It's pretty obvious that Reco's meaning for cheapness was meant to be understood as a reciprocal cost and not as a discount. It might be a legitimate idiom in some parts; who knows. One hears stories of pedants insisting they be paid to carry goods out of the shop because they were labelled "10x cheaper". No way José. Cheers, David.
[toc] | [prev] | [next] | [standalone]
| From | Richard Hector <richard@walnut.gen.nz> |
|---|---|
| Date | 2019-08-11 00:50 +0200 |
| Subject | Re: OT "x times cheaper", was: Re: Server hardware advice. |
| Message-ID | <ytIls-5UO-7@gated-at.bofh.it> |
| In reply to | #212583 |
[Multipart message — attachments visible in raw view] — view raw
On 11/08/19 3:06 AM, David Wright wrote: > On Sat 10 Aug 2019 at 21:19:31 (+1200), Richard Hector wrote: >> On 10/08/19 9:10 PM, deloptes wrote: >>> Richard Hector wrote: >>> >>>> <rant> >>>> Sorry, this usage grates with me. >>>> >>>> $amount cheaper that $price means subtract $amount from $price >>>> >>>> $x times $price means multiply $price by $x >>>> >>>> so "2 times cheaper (than $450)" is: >>>> >>>> $450 - (2 x $450) = -$450. >>> >>> so what multiplied by 2 gives 450? >>> >>> 450 is 100% or 1 >>> 225 is 50% or 1/2 >> >> Right, so 225 is 50% cheaper, or half cheaper. Not twice cheaper. >> >>> perhaps this is the confusion, cause we are using daily language to refer to >>> maths. >> >> Daily language is the problem, yes. I'm not saying my fight is an easy >> one :-) >> >>> In fact I would do it the other way around. >>> >>> initial price x >>> 1xtime x+(1*x) >>> 2xtimes x+(2*x) >>> >>> this gives x=150 >> >> 450 is two times more expensive than 150 (or 200% more than), or three >> times as expensive as 150 (or 300% as expensive). >> 300 is two times as expensive as 150, or 100% more expensive than 150 >> >> We know that these don't work symmetrically; if you have a 50% discount, >> you can't get the original price back by adding 50%, because it's 50% of >> a different number. > > "Expensive" is a dimensional term, like length and time. "Cheap" is in > a different category, like shortness. A 6-inch nail is twice as long > as a 3-inch nail, but one doesn't say the latter is twice as short. Agreed. I prefer to avoid multipliers with inverted dimension terms like that. > But if someone asked for a nail twice as short as this (holding up a > 6-inch nail), you might assume they were a non-native speaker of > English, or you might notice you're almost twice as tall as they are: > ie it's a child. (And it would be polite to offer them a 3-inch > nail. Learning all the categories takes time, and some people might > have slightly different boundaries.) I wouldn't assume that; it's a common usage, even though I consider it wrong :-) A bit like the American habit of saying "I could care less", which also doesn't mean what they mean it to mean :-) > It's pretty obvious that Reco's meaning for cheapness was meant to be > understood as a reciprocal cost and not as a discount. It might be a > legitimate idiom in some parts; who knows. Agreed. And many would consider it a 'legitimate idiom'. I personally consider that from a linguistic and mathematical perspective, it doesn't make sense. > One hears stories of pedants insisting they be paid to carry goods out > of the shop because they were labelled "10x cheaper". No way José. I haven't actually insisted on that, but I've certainly thought it :-) Similarly, one of our local fuel stations has (or had) vouchers that say things like '10c per litre off every litre of fuel' - which also quickly gets into trouble if taken literally :-) Richard
[toc] | [prev] | [next] | [standalone]
| From | mick crane <mick.crane@gmail.com> |
|---|---|
| Date | 2019-08-11 09:50 +0200 |
| Subject | Re: OT "x times cheaper", was: Re: Server hardware advice. |
| Message-ID | <ytQM1-3Wa-9@gated-at.bofh.it> |
| In reply to | #212609 |
On 2019-08-10 23:44, Richard Hector wrote: > On 11/08/19 3:06 AM, David Wright wrote: >> On Sat 10 Aug 2019 at 21:19:31 (+1200), Richard Hector wrote: >>> On 10/08/19 9:10 PM, deloptes wrote: >>>> Richard Hector wrote: >>>> >>>>> <rant> >>>>> Sorry, this usage grates with me. >>>>> >>>>> $amount cheaper that $price means subtract $amount from $price >>>>> >>>>> $x times $price means multiply $price by $x >>>>> >>>>> so "2 times cheaper (than $450)" is: >>>>> >>>>> $450 - (2 x $450) = -$450. >>>> >>>> so what multiplied by 2 gives 450? >>>> >>>> 450 is 100% or 1 >>>> 225 is 50% or 1/2 >>> >>> Right, so 225 is 50% cheaper, or half cheaper. Not twice cheaper. >>> >>>> perhaps this is the confusion, cause we are using daily language to >>>> refer to >>>> maths. >>> >>> Daily language is the problem, yes. I'm not saying my fight is an >>> easy >>> one :-) >>> >>>> In fact I would do it the other way around. >>>> >>>> initial price x >>>> 1xtime x+(1*x) >>>> 2xtimes x+(2*x) >>>> >>>> this gives x=150 >>> >>> 450 is two times more expensive than 150 (or 200% more than), or >>> three >>> times as expensive as 150 (or 300% as expensive). >>> 300 is two times as expensive as 150, or 100% more expensive than 150 >>> >>> We know that these don't work symmetrically; if you have a 50% >>> discount, >>> you can't get the original price back by adding 50%, because it's 50% >>> of >>> a different number. >> >> "Expensive" is a dimensional term, like length and time. "Cheap" is in >> a different category, like shortness. A 6-inch nail is twice as long >> as a 3-inch nail, but one doesn't say the latter is twice as short. > > Agreed. I prefer to avoid multipliers with inverted dimension terms > like > that. > >> But if someone asked for a nail twice as short as this (holding up a >> 6-inch nail), you might assume they were a non-native speaker of >> English, or you might notice you're almost twice as tall as they are: >> ie it's a child. (And it would be polite to offer them a 3-inch >> nail. Learning all the categories takes time, and some people might >> have slightly different boundaries.) > > I wouldn't assume that; it's a common usage, even though I consider it > wrong :-) A bit like the American habit of saying "I could care less", > which also doesn't mean what they mean it to mean :-) > >> It's pretty obvious that Reco's meaning for cheapness was meant to be >> understood as a reciprocal cost and not as a discount. It might be a >> legitimate idiom in some parts; who knows. > > Agreed. And many would consider it a 'legitimate idiom'. I personally > consider that from a linguistic and mathematical perspective, it > doesn't > make sense. > >> One hears stories of pedants insisting they be paid to carry goods out >> of the shop because they were labelled "10x cheaper". No way Jos??. > > I haven't actually insisted on that, but I've certainly thought it :-) > > Similarly, one of our local fuel stations has (or had) vouchers that > say > things like '10c per litre off every litre of fuel' - which also > quickly > gets into trouble if taken literally :-) > > Richard sysadmin types have got *way* too much time on their hands. =o) mick -- Key ID 4BFEBB31
[toc] | [prev] | [next] | [standalone]
| From | Curt <curty@free.fr> |
|---|---|
| Date | 2019-08-11 12:00 +0200 |
| Subject | Re: OT "x times cheaper", was: Re: Server hardware advice. |
| Message-ID | <ytSNP-5b7-1@gated-at.bofh.it> |
| In reply to | #212609 |
On 2019-08-10, Richard Hector <richard@walnut.gen.nz> wrote: > > Similarly, one of our local fuel stations has (or had) vouchers that say > things like '10c per litre off every litre of fuel' - which also quickly > gets into trouble if taken literally :-) You mean that would mean 20c off the second litre and 30c off the third, and so on, or something else? I really don't know. It seems 'every litre of fuel' is where the phrase goes wrong and that part could profitably be deleted, leaving simply: '10c per litre off.' Anyway, whoever wrote that must've been gassed (so to speak). > Richard > > -- “We are all in the gutter, but some of us are looking at the stars.” ― Oscar Wilde, Lady Windermere's Fan
[toc] | [prev] | [next] | [standalone]
| From | John Hasler <jhasler@newsguy.com> |
|---|---|
| Date | 2019-08-09 19:50 +0200 |
| Message-ID | <ythbz-5wK-3@gated-at.bofh.it> |
| In reply to | #212520 |
$430 is way above my budget. "Linksys" and "Wireless" are both negatives. Maybe, if I could get it for $10 at a yard sale... -- John Hasler jhasler@newsguy.com Elmwood, WI USA
[toc] | [prev] | [next] | [standalone]
| From | Steven Mainor <steve@degga.net> |
|---|---|
| Date | 2019-08-10 04:00 +0200 |
| Message-ID | <ytoPL-1P9-1@gated-at.bofh.it> |
| In reply to | #212520 |
[Multipart message — attachments visible in raw view] — view raw
I bought a turris omnia router recently and so far it has worked out pretty well. -- Steven Mainor On August 9, 2019 12:59:34 PM EDT, Reco <recoverym4n@enotuniq.net> wrote: >On Fri, Aug 09, 2019 at 06:16:21PM +0200, deloptes wrote: >> John Hasler wrote: >> >> > Steven Mainor writes: >> > > It looks like there are some ESPRESSOBIN v7s on Amazon right >now. >> > >> > Excellent. When I looked yesterday Amazon said "None available". >I >> > think I'll order one today. The ancient Dell I'm now using as a >> > router/firewall is getting flaky. I've wanted to replace it some >time >> > but I want something ARM-based. This is the first suitable board >I've >> > seen that has dual ethernet. I'll stick it in a box along with the >> > modem, the switch, and a power supply. >> >> This one was very appealing >> >https://www.amazon.de/DMC-Taiwan-Industrial-Networking-Processor/dp/B07T3TWYLJ/ref=sr_1_11?__mk_de_DE=%C3%85M%C3%85%C5%BD%C3%95%C3%91&keywords=acrosser&qid=1561237800&s=gateway&sr=8-11 > >$430 for a router? Surely you're kidding. > >This one is five times cheaper *and* it can run Debian or openwrt: > >https://www.amazon.com/Linksys-Dual-Band-Wireless-Gigabit-WRT1200AC/dp/B00UVN20T0/ref=sr_1_2?keywords=linksys+wrt+1200&qid=1565369861&s=gateway&sr=8-2 > >Reco
[toc] | [prev] | [next] | [standalone]
| From | Steven Mainor <steve@degga.net> |
|---|---|
| Date | 2019-08-07 12:00 +0200 |
| Message-ID | <ysqTE-6kh-5@gated-at.bofh.it> |
| In reply to | #212350 |
[Multipart message — attachments visible in raw view] — view raw
Thanks for the reply. Those seem like options to consider. The pre-orders for the helios4 seem to be sold out for now. -- Steven Mainor On August 7, 2019 2:53:52 AM EDT, Reco <recoverym4n@enotuniq.net> wrote: >On Wed, Aug 07, 2019 at 01:29:21AM -0400, Steven Mainor wrote: >> Hi all, >> >> I'm looking for advice on how to build a home server with a primary >focus on >> security. I plan to run nextcloud and a mail server that will serve 3 >to 5 >> people at most. >> >> My requirements are: >> >> A server setup that can be run with completely open source software >and >> doesn't require any binaries to boot. I don't trust anything closed >source for >> this particular project. >> >> A gigabit ethernet port. >> >> A USB3.0 port or SATA connector to attach storage to. >> >> Enough processor power and ram to run nextcloud and the mail server >from an >> encrypted hard drive (LUKS) efficiently with moderate throughput >saving and >> reading files from nextcloud. > > These fit all your requirements (i.e. it'll run stock buster kernel >without any additional firmware): > >Helios4 - [1]. 4 SATA ports controller attached to PCI-E. >GnuBee - [2]. 6 SATA ports attached to PCI-E. >Odroid HC2 - [3]. Single SATA port, attached to USB bus. > > >> So far I have been looking at single board computers like the ones >listed >> here: https://wiki.debian.org/CheapServerBoxHardware#OSHW > >That list is outdated somewhat. But it gave me good ideas back in the >day. > >Reco > >[1] https://kobol.io/ >[2] http://gnubee.org/ >[3] https://www.hardkernel.com/shop/odroid-hc2-home-cloud-two/
[toc] | [prev] | [next] | [standalone]
| From | Reco <recoverym4n@enotuniq.net> |
|---|---|
| Date | 2019-08-07 13:00 +0200 |
| Message-ID | <ysrPH-6V9-5@gated-at.bofh.it> |
| In reply to | #212361 |
Hi. On Wed, Aug 07, 2019 at 05:58:57AM -0400, Steven Mainor wrote: > Thanks for the reply. Those seem like options to consider. The > pre-orders for the helios4 seem to be sold out for now. They are currently at fourth "campaign", i.e. they're manufacturing a fourth batch. Supply is limited (they produce like a thousand boards per batch), your best bet is a preorder (I got mine at their second "campaign"). Hopefully they do fifth. Reco
[toc] | [prev] | [next] | [standalone]
| From | David Christensen <dpchrist@holgerdanske.com> |
|---|---|
| Date | 2019-08-08 09:10 +0200 |
| Message-ID | <ysKIF-28K-7@gated-at.bofh.it> |
| In reply to | #212344 |
On 8/6/19 10:29 PM, Steven Mainor wrote: > Hi all, > > I'm looking for advice on how to build a home server with a primary focus on > security. Have you considered OpenBSD? Security is their top priority. > I plan to run nextcloud and a mail server that will serve 3 to 5 > people at most. Have you considered a mail hosting provider? The Internet is a war zone and mail servers are prime targets. Are you prepared to fight that battle 24x7? Do you want you home IP that visible? Will your mail users tolerate down-time? I provide e-mail addresses for myself and for family members; I pay professionals to host my mail service. Hosting Nextcloud locally allows you to get at your data at LAN speeds, but: 1. Do you have a good firewall/ router, and know how to use it? 2. Will Nextcloud run in a DMZ? 3. Does Nextcloud require a public IP -- e.g. DNS or dynamic DNS? > My requirements are: > > A server setup that can be run with completely open source software and > doesn't require any binaries to boot. I don't trust anything closed source for > this particular project. I prefer Intel brand motherboards, and Dell products with Intel chips, because the hardware quality is good, support documents are available for many years, and Intel supports FOSS with reference source code. So, a given FOSS OS distribution has the best chance of working OOTB on such hardware without needing binary firmware. That said, the Intel WiFi board I installed in this Dell laptop does require binary firmware. > A gigabit ethernet port. More than one can be very useful. > A USB3.0 port or SATA connector to attach storage to. Internal headers, internal ports, or external ports? SATA DOM? M.2, etc.? I have had issues with FOSS and USB 3.0 ports. USB 2.0 is slower, but more reliable. I have found internal to be more reliable than external. Get at least four internal SATA 6 Gbps ports -- boot disk, optical disk, two data disks (mirrored). I prefer six. > Enough processor power and ram to run nextcloud and the mail server from an > encrypted hard drive (LUKS) efficiently with moderate throughput saving and > reading files from nextcloud. Get a processor with AES-NI support. Get a motherboard that uses ECC RAM and get ECC RAM. Research your OS, file system(s), services, etc., for the amount of RAM needed, and pad according to budget and market price points. I would not get less than 2 @ 4 GB ECC. > I would just build something x86 based but the amd/intel Platform Security > Processor/IME stuff makes me nervous. Do not connect an Ethernet port with a management chip directly to the Internet and/or disable the management chip in the CMOS setup. If the box has more than one Ethernet port and you want the management features, connect the Ethernet port with the management chip to a air-walled administration LAN. > So far I have been looking at single board computers like the ones listed > here: https://wiki.debian.org/CheapServerBoxHardware#OSHW > > I like the OLinuXino A20 LIME2 but I am not sure the processor will be enough > to handle the overhead from an encrypted hard drive. I also don't like that it > is only 32-bit since that will limit the file size nextcloud can handle as I > understand it. > > Is there anything similar to the OLinuXino A20 LIME2 but more powerful or is > there a better option I haven't read about yet? On 8/6/19 11:08 PM, Steven Mainor wrote: > I would like to keep the budget under $500 not including the hard > drive(s) I already have drives. Less is better. On 8/7/19 9:42 AM, Steven Mainor wrote: > For my needs, I doubt anything more than a modern single board > computer is necessary. At least as far as compute power is concerned. I have many doubts about SBC's and SFF computers -- fewer memory, expansion, and I/O slots and ports; fewer choices for processors, memory modules, cases, power supplies, expansion cards, etc.; fewer drive bays; heat issues; long-term vendor support; and the total system cost is always higher than an equivalent tower desktop or server. I built my computers from good quality, standards-compliant, COTS desktop parts for 20+ years. This allowed me to mix-and-match as required when parts died or became obsolete. I still have two 64-bit desktop machines in use, and two 32-bit machines packed in their original boxes in the garage. Then Intel stopped making desktop boards and I wanted ZFS. ZFS wants ECC memory. It was time to migrate to server hardware. I needed to upgrade or replace my SOHO server late last year. My old SOHO server was based on an Antec full tower ATX case and had an Intel D945GNT motherboard, Pentium D processor, and 2 @ 1 GB RAM. One memory slot was bad, preventing installation of 4 @ 1 GB RAM, and dm-crypt without AES-NI cut disk performance in half or more. I replaced the motherboard, CPU, and memory with a used server parts: https://www.ebay.com/itm/Intel-S1200V3RP-Server-Board-Xeon-E3-1225v3-SR1KX-3-20GHz-8GB-Ram-I-O-Shield/123431007451?epid=1941949226&hash=item1cbd0fb4db:g:UuIAAOSwNJlb~bQi:rk:10:pf:0&LH_BIN=1 It works and the performance is amazing, but the case is bulky, the total system requires a fair amount of power, the fans are noisier than I like, and the internal 3.5" drive bays lack vibration isolation (so the HDD heads clatter). Furthermore, I was using old HDD's and several died in as many months. I needed something more modern and reliable. I thought about putting a new Intel single processor server motherboard, Xeon processor, and ECC memory into my Antec Sonata III 500 case, but the cost was prohibitive ($600+). I ended up buying a gently used Dell PowerEdge T30 for US$375. It came with a Xeon E3-1225v5, 1 @ 8 GB ECC, 1 TB SATA HDD, and CD/DVD+-RW. I removed the HDD and added a PC expansion slot 2.5" drive rack, 16 GB SATA SSD, 8 GB ECC, and two 3 TB enterprise SATA HDD's. Total price was less than US$700. It runs FreeBSD 11.2, GELI, ZFS, jails, Samba, and CVS. It is energy efficient, quiet, and very fast. The only downside is that the internal 3.5" drive bays lack vibration isolation. (I need to add sheet rubber to damp the vibrations.) This was the best value proposition I could find. I hope to run it 24x7 for many years. David
[toc] | [prev] | [next] | [standalone]
| From | John Hasler <jhasler@newsguy.com> |
|---|---|
| Date | 2019-08-08 15:20 +0200 |
| Message-ID | <ysQuJ-5H6-7@gated-at.bofh.it> |
| In reply to | #212405 |
Steven Mainor wrote: > I'm looking for advice on how to build a home server with a primary > focus on security. I plan to run nextcloud and a mail server that > will serve 3 to 5 people at most. David Christensen writes: > Have you considered a mail hosting provider? The Internet is a war > zone and mail servers are prime targets. Are you prepared to fight > that battle 24x7? Do you want you home IP that visible? Will your > mail users tolerate down-time? I provide e-mail addresses for myself > and for family members; I pay professionals to host my mail service. Excellent advice. A single Newsguy account will provide all the addresses you need. You can have your domain registrar redirect mail to your Newsguy account so you can still use your domain name. Since you would be paying them (about $5/month) you would have an actual contract, unlike "free" services such as Gmail. -- John Hasler jhasler@newsguy.com Elmwood, WI USA
[toc] | [prev] | [next] | [standalone]
| From | Celejar <celejar@gmail.com> |
|---|---|
| Date | 2019-08-08 15:50 +0200 |
| Message-ID | <ysQXL-5Rz-1@gated-at.bofh.it> |
| In reply to | #212405 |
On Wed, 7 Aug 2019 23:59:44 -0700 David Christensen <dpchrist@holgerdanske.com> wrote: ... > Get at least four internal SATA 6 Gbps ports -- boot disk, optical disk, > two data disks (mirrored). I prefer six. Do most people running servers really want / need an optical disk? As long as the machine can boot via USB, is an optical disk really important for a server? ... > Then Intel stopped making desktop boards and I wanted ZFS. ZFS wants > ECC memory. It was time to migrate to server hardware. My understanding is that ZFS's need / desire for ECC is something of a myth. It's certainly true that many ZFS / FreeNAS *users* have such a need, but the filesystem itself apparently doesn't: https://jrs-s.net/2015/02/03/will-zfs-and-non-ecc-ram-kill-your-data/ Celejar
[toc] | [prev] | [next] | [standalone]
| From | Dan Ritter <dsr@randomstring.org> |
|---|---|
| Date | 2019-08-08 16:30 +0200 |
| Message-ID | <ysRAt-6kT-1@gated-at.bofh.it> |
| In reply to | #212439 |
Celejar wrote: > On Wed, 7 Aug 2019 23:59:44 -0700 > David Christensen <dpchrist@holgerdanske.com> wrote: > > ... > > > Get at least four internal SATA 6 Gbps ports -- boot disk, optical disk, > > two data disks (mirrored). I prefer six. > > Do most people running servers really want / need an optical disk? As > long as the machine can boot via USB, is an optical disk really > important for a server? > > ... > > > Then Intel stopped making desktop boards and I wanted ZFS. ZFS wants > > ECC memory. It was time to migrate to server hardware. > > My understanding is that ZFS's need / desire for ECC is something of a > myth. It's certainly true that many ZFS / FreeNAS *users* have such a > need, but the filesystem itself apparently doesn't: > > https://jrs-s.net/2015/02/03/will-zfs-and-non-ecc-ram-kill-your-data/ To summarize: if you're running ZFS, it can protect you from lots of sources of data corruption. It can't protect you from RAM errors without ECC, so you should opt for ECC if integrity is your goal. None of the other filesystems protect you against RAM errors either, so this is not a special requirement of ZFS. -dsr-
[toc] | [prev] | [next] | [standalone]
| From | Jonas Smedegaard <jonas@jones.dk> |
|---|---|
| Date | 2019-08-08 17:00 +0200 |
| Message-ID | <ysS3v-6w2-21@gated-at.bofh.it> |
| In reply to | #212443 |
[Multipart message — attachments visible in raw view] — view raw
Quoting Dan Ritter (2019-08-08 16:22:07) > Celejar wrote: > > On Wed, 7 Aug 2019 23:59:44 -0700 > > David Christensen <dpchrist@holgerdanske.com> wrote: > > > > ... > > > > > Get at least four internal SATA 6 Gbps ports -- boot disk, optical > > > disk, two data disks (mirrored). I prefer six. > > > > Do most people running servers really want / need an optical disk? > > As long as the machine can boot via USB, is an optical disk really > > important for a server? > > > > ... > > > > > Then Intel stopped making desktop boards and I wanted ZFS. ZFS > > > wants ECC memory. It was time to migrate to server hardware. > > > > My understanding is that ZFS's need / desire for ECC is something of > > a myth. It's certainly true that many ZFS / FreeNAS *users* have > > such a need, but the filesystem itself apparently doesn't: > > > > https://jrs-s.net/2015/02/03/will-zfs-and-non-ecc-ram-kill-your-data/ > > To summarize: if you're running ZFS, it can protect you from lots of > sources of data corruption. It can't protect you from RAM errors > without ECC, so you should opt for ECC if integrity is your goal. > > None of the other filesystems protect you against RAM errors either, > so this is not a special requirement of ZFS. ECC memory is rare among ARM SBCs, but Helios4 uses ECC memory! - Jonas -- * Jonas Smedegaard - idealist & Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ [x] quote me freely [ ] ask before reusing [ ] keep private
[toc] | [prev] | [next] | [standalone]
| From | Reco <recoverym4n@enotuniq.net> |
|---|---|
| Date | 2019-08-08 17:30 +0200 |
| Message-ID | <ysSwx-6VL-3@gated-at.bofh.it> |
| In reply to | #212445 |
Hi. On Thu, Aug 08, 2019 at 04:54:17PM +0200, Jonas Smedegaard wrote: > > > > Then Intel stopped making desktop boards and I wanted ZFS. ZFS > > > > wants ECC memory. It was time to migrate to server hardware. > > > > > > My understanding is that ZFS's need / desire for ECC is something of > > > a myth. It's certainly true that many ZFS / FreeNAS *users* have > > > such a need, but the filesystem itself apparently doesn't: > > > > > > https://jrs-s.net/2015/02/03/will-zfs-and-non-ecc-ram-kill-your-data/ > > > > To summarize: if you're running ZFS, it can protect you from lots of > > sources of data corruption. It can't protect you from RAM errors > > without ECC, so you should opt for ECC if integrity is your goal. > > > > None of the other filesystems protect you against RAM errors either, > > so this is not a special requirement of ZFS. > > ECC memory is rare among ARM SBCs, but Helios4 uses ECC memory! ... with the only problem being the quantity of such RAM. A typical Helios4 board has whopping 2Gb of RAM, which is about 4 times lower than needed for comfortable ZFS usage (assuming that zpool size is measured in terabytes) and a user intends to run something more than a OS kernel and sshd. That estimation deliberately excludes all advanced ZFS features (such as compression, encryption and deduplication). IMO for such RAM sizes it's better to use old trusted MDRAID, LVM, ext4 and a new kid on the block - dm-integrity (all the needed tools are in buster, but some assembly is required). Reco
[toc] | [prev] | [next] | [standalone]
| From | Jonas Smedegaard <jonas@jones.dk> |
|---|---|
| Date | 2019-08-08 18:20 +0200 |
| Message-ID | <ysTiW-7t3-5@gated-at.bofh.it> |
| In reply to | #212449 |
[Multipart message — attachments visible in raw view] — view raw
Quoting Reco (2019-08-08 17:25:02) > Hi. > > On Thu, Aug 08, 2019 at 04:54:17PM +0200, Jonas Smedegaard wrote: > > > > > Then Intel stopped making desktop boards and I wanted ZFS. ZFS > > > > > wants ECC memory. It was time to migrate to server hardware. > > > > > > > > My understanding is that ZFS's need / desire for ECC is > > > > something of a myth. It's certainly true that many ZFS / FreeNAS > > > > *users* have such a need, but the filesystem itself apparently > > > > doesn't: > > > > > > > > https://jrs-s.net/2015/02/03/will-zfs-and-non-ecc-ram-kill-your-data/ > > > > > > To summarize: if you're running ZFS, it can protect you from lots > > > of sources of data corruption. It can't protect you from RAM > > > errors without ECC, so you should opt for ECC if integrity is your > > > goal. > > > > > > None of the other filesystems protect you against RAM errors > > > either, so this is not a special requirement of ZFS. > > > > ECC memory is rare among ARM SBCs, but Helios4 uses ECC memory! > > ... with the only problem being the quantity of such RAM. > > A typical Helios4 board has whopping 2Gb of RAM, which is about 4 > times lower than needed for comfortable ZFS usage (assuming that zpool > size is measured in terabytes) and a user intends to run something > more than a OS kernel and sshd. That estimation deliberately excludes > all advanced ZFS features (such as compression, encryption and > deduplication). > > IMO for such RAM sizes it's better to use old trusted MDRAID, LVM, > ext4 and a new kid on the block - dm-integrity (all the needed tools > are in buster, but some assembly is required). For the record I did not recommend using ZFS on low-end hardware. The OP asked for advice in buying low-end ARM-based hardware for use as server, and I pointed out that one ARM SBC (likely the only relatively cheap one) is known to use ECC memory - which (as the previous poster pointed out) is interesting _independently_ of choice of filesystem. Personally I use ext4 with journaling enabled, on either conventional rotating disks, SSDs, or sdcards (no RAID involved). - Jonas -- * Jonas Smedegaard - idealist & Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ [x] quote me freely [ ] ask before reusing [ ] keep private
[toc] | [prev] | [next] | [standalone]
| From | Steven Mainor <steve@degga.net> |
|---|---|
| Date | 2019-08-08 22:50 +0200 |
| Message-ID | <ysXwd-1AN-1@gated-at.bofh.it> |
| In reply to | #212451 |
[Multipart message — attachments visible in raw view] — view raw
So is the general consensus that there are no modern SBCs powerful enough to run nextcloud on (apache, mariadb, php) or a mail server (typical postfix, dovecot, opendkim, SpamAssassin etc... ) for a handful of people? That seems hard to believe. -- Steven Mainor On August 8, 2019 12:14:23 PM EDT, Jonas Smedegaard <jonas@jones.dk> wrote: >Quoting Reco (2019-08-08 17:25:02) >> Hi. >> >> On Thu, Aug 08, 2019 at 04:54:17PM +0200, Jonas Smedegaard wrote: >> > > > > Then Intel stopped making desktop boards and I wanted ZFS. >ZFS >> > > > > wants ECC memory. It was time to migrate to server hardware. >> > > > >> > > > My understanding is that ZFS's need / desire for ECC is >> > > > something of a myth. It's certainly true that many ZFS / >FreeNAS >> > > > *users* have such a need, but the filesystem itself apparently >> > > > doesn't: >> > > > >> > > > >https://jrs-s.net/2015/02/03/will-zfs-and-non-ecc-ram-kill-your-data/ >> > > >> > > To summarize: if you're running ZFS, it can protect you from lots > >> > > of sources of data corruption. It can't protect you from RAM >> > > errors without ECC, so you should opt for ECC if integrity is >your >> > > goal. >> > > >> > > None of the other filesystems protect you against RAM errors >> > > either, so this is not a special requirement of ZFS. >> > >> > ECC memory is rare among ARM SBCs, but Helios4 uses ECC memory! >> >> ... with the only problem being the quantity of such RAM. >> >> A typical Helios4 board has whopping 2Gb of RAM, which is about 4 >> times lower than needed for comfortable ZFS usage (assuming that >zpool >> size is measured in terabytes) and a user intends to run something >> more than a OS kernel and sshd. That estimation deliberately excludes > >> all advanced ZFS features (such as compression, encryption and >> deduplication). >> >> IMO for such RAM sizes it's better to use old trusted MDRAID, LVM, >> ext4 and a new kid on the block - dm-integrity (all the needed tools >> are in buster, but some assembly is required). > >For the record I did not recommend using ZFS on low-end hardware. > >The OP asked for advice in buying low-end ARM-based hardware for use as > >server, and I pointed out that one ARM SBC (likely the only relatively >cheap one) is known to use ECC memory - which (as the previous poster >pointed out) is interesting _independently_ of choice of filesystem. > >Personally I use ext4 with journaling enabled, on either conventional >rotating disks, SSDs, or sdcards (no RAID involved). > > > - Jonas > >-- > * Jonas Smedegaard - idealist & Internet-arkitekt > * Tlf.: +45 40843136 Website: http://dr.jones.dk/ > > [x] quote me freely [ ] ask before reusing [ ] keep private
[toc] | [prev] | [next] | [standalone]
Page 4 of 5 — ← Prev page 1 2 3 [4] 5 Next page →
Back to top | Article view | linux.debian.user
csiph-web