A GREAT MANY people can tell the same story. SORBS was half-automated. Getting on any of their lists was simple -- a SINGLE email is all it takes. Getting off the list is not automated in ANY way. It takes a human being -- and by all accounts, there was only ever one human -- to remove you. The removal process was never speedy and usually involved a ransom in the form of a donation to an ever dwindling number of charities. (no respectable charity would have anything to do with this crap.)
The internet is a constantly changing landscape. SORBS was always a stick in the mud refusing to face the realities of the changing landscape. Spammers move around. Address space gets reassigned. A single email is not a valid reason to blacklist an address for ever more.
The servers, software, database, "reputation", etc. Basically, an anti-spam operation on wheels. ('tho I don't recommend rolling fully loaded racks around. A fully loaded 43U rack can weigh nearly a ton.)
... who have random mailservers listed in SORBS' spamtrap list. The day my mailserver dropped a message from gmail is the day I stopped using SORBS. This was cast in cement by their bitchy response to my inquiry... "[gmail] doesn't do outbound content filtering, so f*** 'em."
Having worked for a former ISP of Jermey Janes, lemme just say, it's impossible to educate sales people. They care about their commission, and nothing else. When I checked, they had, in fact, had several connections with us for several years -- all under different names, to different addresses. But T1's were just too small for any real volume. When they ordered a DS3 and colo space, they screwed up and sent out 100,000 "template" spam messages and abuse@ received "many" complaints -- and the f'ing morons included some of our addresses as well (spamming your own ISP... priceless.) They were disconnected about an hour after their circuit was turned up.:-) [That's when I went hunting.] We terminated their connectivity as per the terms of our contract, but they remained as a colo customer -- $175/hr "escorted access" on top of colo was "madd money". (almost made up for PSI going out of business.)
SPAMHAUS was informed of all their address blocks -- and a few aliases/shell corps. they didn't know about. They were even told of the/20 we were allocated for them -- ARIN refused to give them anything -- before it even appeared in whois.:-)
Lemme guess, there answer was "spamtrap"? For which they will give zero evidence and demand payment to get removed. (That's extortion, btw. But since "we never touch the money" it's not illegal.) All it takes is ONE MESSAGE, EVER, to land in the spamtrap list; listings never expire, and you have to make a "donation" to get delisted. Most interesting is the number of charities refusing to be associated with the SORBS extortion racket.
While I doubt there was much intentional (mis)listings, there certainly were a number of listing made as a strong-arm tactic. ('spam friendly XXX' falls in that bucket.)
SusClientID not SID. WSUS identifies a machine based on a self-generated ID completely independent of the machine/domain SID. It won't change until the registry key is deleted.
CD/DVD's cost pennies to make and are pressed by the thousands per hour. VHS tapes take a lot longer and cost a lot more. Rental chains got the "first runs" at insanely higher prices ($100 and $150 PER TAPE were common) before there were quantities available for retail purchase. As such, rentals were a massing source of income. That market is gone, now. They'd rather get $19.95 for a DVD that cost $0.10 to make.
Microsoft does actually put "everything" online. Maybe you've heard of MSDN? However, most people don't buy an MSDN subscription nor do they buy their copy of Windows directly from Microsoft.
Just because the CD/DVD is infested with DRM does not mean it's not easily circumvented. (because it is.)
Interesting theory, but if you'll recall, they've used that same effect for hyperspace transitions on goa'uld motherships.
It'll be interesting to see what BS they make up to explain it. It appears that ship is moving at unimaginable speeds... they jumped from one galaxy to another in a few hours (assuming they were in a galaxy when they gated aboard.)
That's muddy waters... Does downloading a demo ("free") app constitute a "business relationship"? As for telemarketing calls to cellphones, it's certainly despised, but I don't think it's illegal these days -- for starters, it's impossible to know the number you're dialing is a cellphone, or has been directed to a cellphone. The days when an NPANXX could tell you a location and service provider are long past. (any number can be assigned to anyone, anywhere.)
As a general rule: No, no they cannot. I've been at a number of (stupid) companies where they ask the employees to do "self evaluations", which then goes up the tree for signatures. The employees hate wasting their time on the bullshit, because they're never questioned and very rarely read. In other places, the manager(s) do the evals -- which amounts to filling in a form. Problem there is they aren't fully qualified to evaluate anyone's work, and most times don't fully understand what their minions do.
Modern electronics aren't "wave soldered" anymore. For starters, lead is "four letter word" and no longer allowed on Earth. Today's electronics are too small and delicate to be run through an 800 degree pool -- it'd melt everything and short all the tiny contacts. They use silkscreened solder paste and low temperature IR ovens.
Better not let them see that or they'll fuck it up everywhere. Less than 24hrs after pointing out that non-logged in users and users of "other" browsers (safari for one) were still presented the old interface, they "corrected" that. Every browser now gets the same javascript bloatware.
No. The real problem is that the marketing types are "morons". They aren't engineers and thus have zero understanding of the complexities and inherent brokenness of the shit they dream up. "I want the combo boxes to match the look and feel of the rest of the page. And for everything to look the same on every browser." I don't have to work for eBay to know those words were said. If your website won't even display correctly without a ton of javascript, you are an idiot and should be executed for the public good.
It's not Opera's fault at all. It's 100% eBay's screwed up javascript page rendering crap that doesn't work correctly outside IE (and sometimes not even there.) Safari, Chrome, and Firefox have all had issues with the eBay Beta interface.
That BS is not limited to Opera. And I've complained about it repeatedly. They don't give a flying f***. If you aren't using the latest IE, they don't want to hear it. (And when you are, it's not their problem; it's something in *your* browser that's messed up.)
What everyone is now being forced to use, used be called "The eBay Beta". It's been around for months -- maybe years. And I've been bitching about the horrible browser compatibility, extreme abuse of javascript that places supercomputer loads on browsers, etc. since DAY ONE. They don't listen; and they don't test jack s***. What used to be a combo box handled 100% by the browser is now a complex set of images and javascript -- just to make some asshole in marketing happy by "mak[ing] everything look the same". There is ZERO f'ing reason for there to be soooooo g**da** much javascript for something as simple as a web page. (They do it to make it as near impossible as they can to block all the crap ads they plaster everywhere.)
Nobody has ever been happy with the feedback system. Too many people abuse it and extort people through it. If I leave a negative about a seller, they will almost without fail leave negative for me. (and v.v.)
It would be interesting to see actual performance numbers. On paper, based on the performance of my own sil3124 and 1TB drives, if it's doing nothing else, it should be able to rebuild in just under 6 days. However, given that it's under load, there are port multipliers, and the 3124 shares bandwidth across all of it's ports, it might take twice that long.
(PCI-X vs. PCIe makes little difference. The drives are nowhere near as fast as either bus.)
Actually, it's not that bad. Half a day, maybe. They didn't necessarily build them to be fast -- they use port multipliers, so there's 80% of the performance gone.
However, the point still stands... they *do* have multiple layers of data integrity, both within each unit and across the data center.
And for a contrasting story... Cisco Support. A few years ago, while working for a telco, we had a Cisco 2920 fail -- it's a fixed configuration 2 slot Cat 5000. Cisco stopped making those things a thousand years ago, but they kept taking our money for hardware support contracts (we had a dozen of them) which means they're legally required to have spares. After several hours working with a tech -- removing cards that aren't supposed to be removable, it was determined the backplane was bad; so no shipping me parts from a 5500 (which I already had.) I could hear the guy shaking his head over the phone:-) "Lemme check around and see if I can find one. I'll call you back." It took a while, but he found one and had it there the next morning -- which, short of buying it a seat on a plane, was as fast as it could get here. The following Monday, they canceled (and refunded) our service contracts on those things. heh.
Actually, they did. You fail as a geek:-) They run RAID6 across blocks of 15 drives. As long as they scrub each array regularly, they'll detect and correct any corruption or bad disks before it becomes an issue.
ECC RAM is a bit of an unnecessary expense. Bit errors in RAM are exceedingly rare. I have many (MANY) servers with ECC memory. Over nearly a century of total CPU time, none of them have ever reported ECC errors. (even after over heating the hell out of one of them.)
And if you had it setup to send you those email alerts, you would've known about the failure before they did.
What BackBlaze leaves out is the quality of equipment from the "expensive" vendors. They built their toy with the cheapest shit they could find -- except for the power switch, that was $30! NetApp, EMC, et. al. don't go down to Best Buy and get a shopping cart full of drives (I've done that, for the record.) The big boys use much more stable and reliable SCSI hardware -- Fibre Channel and SAS. Lemme tell ya', today's SATA/IDE drives are complete shit compared to those FC/SAS drives. I personally have 10 year old FC drives still in service today. I don't have a single IDE drive that has lasted more than about 3 years; and if you leave one sitting on a shelf, your data is decaying by the nanosecond.
There's a very good reason why some things are cheap and some things are expensive.
If their infrastructure can support multiple drive failures per day, then it might be worth it. I'm not in that business, so I've not taken a detailed look at the numbers. I'm an admin who wants my data to be there when I wake up every morning. And I don't want the headache of having to keep multiple copies on multiple servers for the times drives and servers fail.
Slashdot Mirror
A GREAT MANY people can tell the same story. SORBS was half-automated. Getting on any of their lists was simple -- a SINGLE email is all it takes. Getting off the list is not automated in ANY way. It takes a human being -- and by all accounts, there was only ever one human -- to remove you. The removal process was never speedy and usually involved a ransom in the form of a donation to an ever dwindling number of charities. (no respectable charity would have anything to do with this crap.)
The internet is a constantly changing landscape. SORBS was always a stick in the mud refusing to face the realities of the changing landscape. Spammers move around. Address space gets reassigned. A single email is not a valid reason to blacklist an address for ever more.
The servers, software, database, "reputation", etc. Basically, an anti-spam operation on wheels. ('tho I don't recommend rolling fully loaded racks around. A fully loaded 43U rack can weigh nearly a ton.)
450k isn't that much money as these things go. Still, that's a lot of folding money in one's pocket.
... who have random mailservers listed in SORBS' spamtrap list. The day my mailserver dropped a message from gmail is the day I stopped using SORBS. This was cast in cement by their bitchy response to my inquiry... "[gmail] doesn't do outbound content filtering, so f*** 'em."
Having worked for a former ISP of Jermey Janes, lemme just say, it's impossible to educate sales people. They care about their commission, and nothing else. When I checked, they had, in fact, had several connections with us for several years -- all under different names, to different addresses. But T1's were just too small for any real volume. When they ordered a DS3 and colo space, they screwed up and sent out 100,000 "template" spam messages and abuse@ received "many" complaints -- and the f'ing morons included some of our addresses as well (spamming your own ISP... priceless.) They were disconnected about an hour after their circuit was turned up. :-) [That's when I went hunting.] We terminated their connectivity as per the terms of our contract, but they remained as a colo customer -- $175/hr "escorted access" on top of colo was "madd money". (almost made up for PSI going out of business.)
SPAMHAUS was informed of all their address blocks -- and a few aliases/shell corps. they didn't know about. They were even told of the /20 we were allocated for them -- ARIN refused to give them anything -- before it even appeared in whois. :-)
Lemme guess, there answer was "spamtrap"? For which they will give zero evidence and demand payment to get removed. (That's extortion, btw. But since "we never touch the money" it's not illegal.) All it takes is ONE MESSAGE, EVER, to land in the spamtrap list; listings never expire, and you have to make a "donation" to get delisted. Most interesting is the number of charities refusing to be associated with the SORBS extortion racket.
While I doubt there was much intentional (mis)listings, there certainly were a number of listing made as a strong-arm tactic. ('spam friendly XXX' falls in that bucket.)
SusClientID not SID. WSUS identifies a machine based on a self-generated ID completely independent of the machine/domain SID. It won't change until the registry key is deleted.
CD/DVD's cost pennies to make and are pressed by the thousands per hour. VHS tapes take a lot longer and cost a lot more. Rental chains got the "first runs" at insanely higher prices ($100 and $150 PER TAPE were common) before there were quantities available for retail purchase. As such, rentals were a massing source of income. That market is gone, now. They'd rather get $19.95 for a DVD that cost $0.10 to make.
Microsoft does actually put "everything" online. Maybe you've heard of MSDN? However, most people don't buy an MSDN subscription nor do they buy their copy of Windows directly from Microsoft.
Just because the CD/DVD is infested with DRM does not mean it's not easily circumvented. (because it is.)
Interesting theory, but if you'll recall, they've used that same effect for hyperspace transitions on goa'uld motherships.
It'll be interesting to see what BS they make up to explain it. It appears that ship is moving at unimaginable speeds... they jumped from one galaxy to another in a few hours (assuming they were in a galaxy when they gated aboard.)
That's muddy waters... Does downloading a demo ("free") app constitute a "business relationship"? As for telemarketing calls to cellphones, it's certainly despised, but I don't think it's illegal these days -- for starters, it's impossible to know the number you're dialing is a cellphone, or has been directed to a cellphone. The days when an NPANXX could tell you a location and service provider are long past. (any number can be assigned to anyone, anywhere.)
As a general rule: No, no they cannot. I've been at a number of (stupid) companies where they ask the employees to do "self evaluations", which then goes up the tree for signatures. The employees hate wasting their time on the bullshit, because they're never questioned and very rarely read. In other places, the manager(s) do the evals -- which amounts to filling in a form. Problem there is they aren't fully qualified to evaluate anyone's work, and most times don't fully understand what their minions do.
Modern electronics aren't "wave soldered" anymore. For starters, lead is "four letter word" and no longer allowed on Earth. Today's electronics are too small and delicate to be run through an 800 degree pool -- it'd melt everything and short all the tiny contacts. They use silkscreened solder paste and low temperature IR ovens.
(But, yes, I loved watching wave soldering rigs.)
Better not let them see that or they'll fuck it up everywhere. Less than 24hrs after pointing out that non-logged in users and users of "other" browsers (safari for one) were still presented the old interface, they "corrected" that. Every browser now gets the same javascript bloatware.
No. The real problem is that the marketing types are "morons". They aren't engineers and thus have zero understanding of the complexities and inherent brokenness of the shit they dream up. "I want the combo boxes to match the look and feel of the rest of the page. And for everything to look the same on every browser." I don't have to work for eBay to know those words were said. If your website won't even display correctly without a ton of javascript, you are an idiot and should be executed for the public good.
That's the difference between "engineer" and "marketing". Engineers care that things actually function. Marketing only cares that it looks good.
It's not Opera's fault at all. It's 100% eBay's screwed up javascript page rendering crap that doesn't work correctly outside IE (and sometimes not even there.) Safari, Chrome, and Firefox have all had issues with the eBay Beta interface.
That BS is not limited to Opera. And I've complained about it repeatedly. They don't give a flying f***. If you aren't using the latest IE, they don't want to hear it. (And when you are, it's not their problem; it's something in *your* browser that's messed up.)
What everyone is now being forced to use, used be called "The eBay Beta". It's been around for months -- maybe years. And I've been bitching about the horrible browser compatibility, extreme abuse of javascript that places supercomputer loads on browsers, etc. since DAY ONE. They don't listen; and they don't test jack s***. What used to be a combo box handled 100% by the browser is now a complex set of images and javascript -- just to make some asshole in marketing happy by "mak[ing] everything look the same". There is ZERO f'ing reason for there to be soooooo g**da** much javascript for something as simple as a web page. (They do it to make it as near impossible as they can to block all the crap ads they plaster everywhere.)
Nobody has ever been happy with the feedback system. Too many people abuse it and extort people through it. If I leave a negative about a seller, they will almost without fail leave negative for me. (and v.v.)
It would be interesting to see actual performance numbers. On paper, based on the performance of my own sil3124 and 1TB drives, if it's doing nothing else, it should be able to rebuild in just under 6 days. However, given that it's under load, there are port multipliers, and the 3124 shares bandwidth across all of it's ports, it might take twice that long.
(PCI-X vs. PCIe makes little difference. The drives are nowhere near as fast as either bus.)
Actually, it's not that bad. Half a day, maybe. They didn't necessarily build them to be fast -- they use port multipliers, so there's 80% of the performance gone.
However, the point still stands... they *do* have multiple layers of data integrity, both within each unit and across the data center.
And for a contrasting story... Cisco Support. A few years ago, while working for a telco, we had a Cisco 2920 fail -- it's a fixed configuration 2 slot Cat 5000. Cisco stopped making those things a thousand years ago, but they kept taking our money for hardware support contracts (we had a dozen of them) which means they're legally required to have spares. After several hours working with a tech -- removing cards that aren't supposed to be removable, it was determined the backplane was bad; so no shipping me parts from a 5500 (which I already had.) I could hear the guy shaking his head over the phone :-) "Lemme check around and see if I can find one. I'll call you back." It took a while, but he found one and had it there the next morning -- which, short of buying it a seat on a plane, was as fast as it could get here. The following Monday, they canceled (and refunded) our service contracts on those things. heh.
Actually, they did. You fail as a geek :-) They run RAID6 across blocks of 15 drives. As long as they scrub each array regularly, they'll detect and correct any corruption or bad disks before it becomes an issue.
ECC RAM is a bit of an unnecessary expense. Bit errors in RAM are exceedingly rare. I have many (MANY) servers with ECC memory. Over nearly a century of total CPU time, none of them have ever reported ECC errors. (even after over heating the hell out of one of them.)
And if you had it setup to send you those email alerts, you would've known about the failure before they did.
What BackBlaze leaves out is the quality of equipment from the "expensive" vendors. They built their toy with the cheapest shit they could find -- except for the power switch, that was $30! NetApp, EMC, et. al. don't go down to Best Buy and get a shopping cart full of drives (I've done that, for the record.) The big boys use much more stable and reliable SCSI hardware -- Fibre Channel and SAS. Lemme tell ya', today's SATA/IDE drives are complete shit compared to those FC/SAS drives. I personally have 10 year old FC drives still in service today. I don't have a single IDE drive that has lasted more than about 3 years; and if you leave one sitting on a shelf, your data is decaying by the nanosecond.
There's a very good reason why some things are cheap and some things are expensive.
If their infrastructure can support multiple drive failures per day, then it might be worth it. I'm not in that business, so I've not taken a detailed look at the numbers. I'm an admin who wants my data to be there when I wake up every morning. And I don't want the headache of having to keep multiple copies on multiple servers for the times drives and servers fail.
And they too would be incompetent in Childs' eye. "I built the network, so noone can understand it better than me." Arrogant? Yes. Correct? No.