only because DH objected. The point of my comment was that had they not objected there was an issue with the initial request that could have been abused.
Of course, but the precedent of just handing over such data is a *very* bad one. DH wasn't just protecting the current crop of website visitors, they're protecting the entire US population of website visitors. It's not a stretch to think that some other agency, that is based in a slightly off-white house-type building, would get a copy quietly and trawl that for political opponents.
Yup. I bought a Samsung Smart TV shortly after my divorce left me without pretty much anything, a little 24" model. Got it (and used the smart feature) specifically for netflix, effectively as an integrated STB. Based on the *forced* updates, just how long they took, the annoying habit to change UI layout for no good reason, and that it would prevent you from using the TV at all while they were running, I disabled its network connection. Switched to a chromecast and never looked back.
I think 7 was an improvement on XP64, much how XP was an improvement on 2000. I would like to see WinXP &&|| Win7 open sourced. Vista too, specifically for the diffs to XP and 7 so we can see what *not* to do!
Well, TBH the central banks are *not* the ones in power. Rothchild family has control over simply obscene amounts of money... If they decide you're not moving your money it doesn't matter if you're a nobody or a country, it's going to be damn hard to move the money.
I have T-Mobile and have my account set to *require* in store identification to move phone number. I tested it and so far they've not let me move my phone number away to my spare phone, replying only that "I'm sorry sir but your account is very specific that you must go into our store and provide proper identification and pin before you can move your number. I would be happy to provide you a temporary number until you can get to a retail location."
So, they could still rack up charges on my account with a temp number I suppose, but at least can't redirect my actual number.
My truck was burgled and I lost a checkbook. Called my bank that morning and they basically said the same thing: We're cancelling all your checks for you now, but be sure to file a police report, even though they won't do anything, just so when someone tries to fraudulently pass off your checks you have a report that they were stolen from *before* the attempt was made.
Also in OP's case I would consider a lawsuit against Verizon.
coinbase uses the authy app (similar to RSA) such that if you lose your phone number you are still not locked out of your account (and they can't get in). BUT if you have no other account recovery options in place and your phone goes titsup you're in trouble!
That is the crux of the problem, users want both super ease of use and high security; they are quite often opposite sides of the coin.
Drives don't spin down, and I have CPU currently set to no C sleep states, so NOPs don't cause core frequency to drop. I shot for absolutely most stable config, and am changing one thing at a time per month as I work on lowering power.
Also, the FusionIOs have a funky power configuration, I think that they are doing a lot of background stuff that I could configure more conservatively... but honestly getting them up and running was enough of a pain that I'm afraid to touch them. I definitely would advise against using them in future builds, and just using the newer Intel cards and paying the price premium if you really want that kind of speed. I mean I got the pair for $340, so killer deal, but... *pain*
I bet building your own would have taken less time than what you did looking for a "cheap, fast, power efficient, non-proprietary" unicorn.
Unicorn is right, and you left out "configurable".
I do understand his desire for the unicorn solution though. I rebuilt my NAS *three* times while learning all the ins and outs of what I wanted and how best to accomplish it. Upside, those 6x6TB drives got a thorough burn-in test!
Dunno what that AC was thinking... My real world experience: It is fast enough that I have not had any issues, I run two pools, one for bulk storage, one for multiple stream access. Bulk storage is on 6TB NAS drives in mirror sets, one WD one ST and one ANY per set. Two such sets in one Zpool (so 6 drives for 12 TB of storage) high reliability, high enough access as any set can effectively dedicate a drive per read stream. Obviously serving a 4th read stream would possibly cause issues quickly, and for sure a 7th stream would; depending on file location. For the pool that actually is expected to be serving multiple streams it's SSD based (pair of refurb FusionIO 640G PCIe SSDs in single 1.2TB pool) These are backed up to the other larger pool regularly (as part of the scrub cronjob).
Use case: Host for two iSCSI targets for VMWare ESXi system Streaming server for PLEX front-end (mounted as NFS) host for minecraft server (mounted as NFS) general filer for my home network. At any point the system may be serving one plex stream and two iSCSI endpoints (PLEX is on the shared GigE, iSCSI each have their own) from the 12TB pool and serving file requests for all NFS mounts other than PLEX from the SSD.
Now power: Excellent, likely close to what the integrated systems use... about 250w under load/150W idle. (6 drives, 2 Fusion IO, one onboard lan, one x4 2 port GigE) onboard SATA, integrated GFX, 16 gig ECC ram, i5 2 core cpu).
My solution was born from not finding anything commercial that was even close to the featureset I wanted at a price that was palatable. I'm sure it's not as perfect as a high end commercial system that's been properly tuned, but it meets my needs for about 1/4 the cost.
not that they have the code tables memorized! That'd be crazy.
I mean more like hey this system is outputting funky video and knowing the things to check (and report in ticket): display panel FW blob is for correct chipset what BAR registers to get dumps from for video
Things like that, vs just sending in a ticket like: Video on vendor foo product bar is acting funny, what causes that? can you tell me what FW ver I should be on? (without telling you what step and option set is in this system) have you seen this before? sometimes without even the output of the code reader:(
one bad cell in the pack can render the entire pack useless. TFA says they test each cell through a discharge/charge cycle, my guess being that they cull the horrible ones, and bin sort the rest.
IIRC every parallel element in a Li pack should be monitored and individually controlled, correct?
so I can have 100 cells in 10 strings of 10 and would need 10 controllers* to monitor said pack.
Then the parallel pack itself can be treated somewhat simply as a larger battery equivalent (e.g. if each of the hypothetical strings is 1Ah then I can treat the pack as a single 10Ah battery). The danger comes into play when cells are in parallel without protection, as then the stronger cell can actually damage the weaker cell in boundary conditions, leading to over depletion and over charging situations.
*in this case controller is specifically: LV cutoff to prevent over depletion of string Charge current limiting HV cutoff possibly max current drain too.
good systems engineers know all the default and most common fault conditions of the systems they build/maintain. *Great* systems engineers know all that and most of the less common fault conditions *and* know how to provide accurate debug data & questions to the developers such that an answer has a high probability of being the one you needed. a.k.a. asking the right question.
It was something that I fought with all the damn time when I was at my former employer. SIs asked shotgun questions and provided no debug data, then bitched when we didn't answer the question "correctly". The good ones usually got their answers in half a day because they gave a debug dump, clear description of the problem & what was expected, and dead perfect reproduction steps.
I have a 36 hour cron job rsync that runs on my critical work. Everything else is weekly. Plus a pair of local drives that take turns being on the system every other week.
Slashdot Mirror
only because DH objected. The point of my comment was that had they not objected there was an issue with the initial request that could have been abused.
Of course, but the precedent of just handing over such data is a *very* bad one.
DH wasn't just protecting the current crop of website visitors, they're protecting the entire US population of website visitors.
It's not a stretch to think that some other agency, that is based in a slightly off-white house-type building, would get a copy quietly and trawl that for political opponents.
So you're saying they're Santa Claus?
Yup. I bought a Samsung Smart TV shortly after my divorce left me without pretty much anything, a little 24" model. Got it (and used the smart feature) specifically for netflix, effectively as an integrated STB. Based on the *forced* updates, just how long they took, the annoying habit to change UI layout for no good reason, and that it would prevent you from using the TV at all while they were running, I disabled its network connection. Switched to a chromecast and never looked back.
I think 7 was an improvement on XP64, much how XP was an improvement on 2000.
I would like to see WinXP &&|| Win7 open sourced. Vista too, specifically for the diffs to XP and 7 so we can see what *not* to do!
Well, TBH the central banks are *not* the ones in power. Rothchild family has control over simply obscene amounts of money... If they decide you're not moving your money it doesn't matter if you're a nobody or a country, it's going to be damn hard to move the money.
indeed, AC was right. With just how bad that hurt me to read...
nothing I've seen in that price range had the features I wanted like iSCSI target, multiple LAN, etc.
I realise I wanted more a "baby SAN" than a home NAS, but hey. I also got to learn a ton and increase my skill base for prospective employers.
I have T-Mobile and have my account set to *require* in store identification to move phone number.
I tested it and so far they've not let me move my phone number away to my spare phone, replying only that "I'm sorry sir but your account is very specific that you must go into our store and provide proper identification and pin before you can move your number. I would be happy to provide you a temporary number until you can get to a retail location."
So, they could still rack up charges on my account with a temp number I suppose, but at least can't redirect my actual number.
My truck was burgled and I lost a checkbook.
Called my bank that morning and they basically said the same thing:
We're cancelling all your checks for you now, but be sure to file a police report, even though they won't do anything, just so when someone tries to fraudulently pass off your checks you have a report that they were stolen from *before* the attempt was made.
Also in OP's case I would consider a lawsuit against Verizon.
more detail about Authy issue please... I use it!
coinbase uses the authy app (similar to RSA) such that if you lose your phone number you are still not locked out of your account (and they can't get in). BUT if you have no other account recovery options in place and your phone goes titsup you're in trouble!
That is the crux of the problem, users want both super ease of use and high security; they are quite often opposite sides of the coin.
Drives don't spin down, and I have CPU currently set to no C sleep states, so NOPs don't cause core frequency to drop.
I shot for absolutely most stable config, and am changing one thing at a time per month as I work on lowering power.
Also, the FusionIOs have a funky power configuration, I think that they are doing a lot of background stuff that I could configure more conservatively... but honestly getting them up and running was enough of a pain that I'm afraid to touch them. I definitely would advise against using them in future builds, and just using the newer Intel cards and paying the price premium if you really want that kind of speed. I mean I got the pair for $340, so killer deal, but... *pain*
I bet building your own would have taken less time than what you did looking for a "cheap, fast, power efficient, non-proprietary" unicorn.
Unicorn is right, and you left out "configurable".
I do understand his desire for the unicorn solution though. I rebuilt my NAS *three* times while learning all the ins and outs of what I wanted and how best to accomplish it.
Upside, those 6x6TB drives got a thorough burn-in test!
Dunno what that AC was thinking...
My real world experience: It is fast enough that I have not had any issues, I run two pools, one for bulk storage, one for multiple stream access.
Bulk storage is on 6TB NAS drives in mirror sets, one WD one ST and one ANY per set. Two such sets in one Zpool (so 6 drives for 12 TB of storage) high reliability, high enough access as any set can effectively dedicate a drive per read stream. Obviously serving a 4th read stream would possibly cause issues quickly, and for sure a 7th stream would; depending on file location.
For the pool that actually is expected to be serving multiple streams it's SSD based (pair of refurb FusionIO 640G PCIe SSDs in single 1.2TB pool) These are backed up to the other larger pool regularly (as part of the scrub cronjob).
Use case:
Host for two iSCSI targets for VMWare ESXi system
Streaming server for PLEX front-end (mounted as NFS)
host for minecraft server (mounted as NFS)
general filer for my home network.
At any point the system may be serving one plex stream and two iSCSI endpoints (PLEX is on the shared GigE, iSCSI each have their own) from the 12TB pool and serving file requests for all NFS mounts other than PLEX from the SSD.
Now power: Excellent, likely close to what the integrated systems use... about 250w under load/150W idle.
(6 drives, 2 Fusion IO, one onboard lan, one x4 2 port GigE) onboard SATA, integrated GFX, 16 gig ECC ram, i5 2 core cpu).
My solution was born from not finding anything commercial that was even close to the featureset I wanted at a price that was palatable. I'm sure it's not as perfect as a high end commercial system that's been properly tuned, but it meets my needs for about 1/4 the cost.
EE / MarketingCommunications? :p
That's the fast track to psychopathy and multiple personality disorder
not that they have the code tables memorized! That'd be crazy.
I mean more like hey this system is outputting funky video and knowing the things to check (and report in ticket):
display panel
FW blob is for correct chipset
what BAR registers to get dumps from for video
Things like that, vs just sending in a ticket like: :(
Video on vendor foo product bar is acting funny, what causes that?
can you tell me what FW ver I should be on? (without telling you what step and option set is in this system)
have you seen this before?
sometimes without even the output of the code reader
one bad cell in the pack can render the entire pack useless.
TFA says they test each cell through a discharge/charge cycle, my guess being that they cull the horrible ones, and bin sort the rest.
IIRC every parallel element in a Li pack should be monitored and individually controlled, correct?
so I can have 100 cells in 10 strings of 10 and would need 10 controllers* to monitor said pack.
Then the parallel pack itself can be treated somewhat simply as a larger battery equivalent (e.g. if each of the hypothetical strings is 1Ah then I can treat the pack as a single 10Ah battery). The danger comes into play when cells are in parallel without protection, as then the stronger cell can actually damage the weaker cell in boundary conditions, leading to over depletion and over charging situations.
*in this case controller is specifically:
LV cutoff to prevent over depletion of string
Charge current limiting
HV cutoff
possibly max current drain too.
Nas4Free and a relatively low power i5 ECC system with drive bays turned out *cheaper* than a small standards compliant NAS.
sad but true.
I have optical/rca to the pre-amp, but the pre-am to amp connections are all over balanced lines, specifically, XLR.
good systems engineers know all the default and most common fault conditions of the systems they build/maintain.
*Great* systems engineers know all that and most of the less common fault conditions *and* know how to provide accurate debug data & questions to the developers such that an answer has a high probability of being the one you needed.
a.k.a. asking the right question.
It was something that I fought with all the damn time when I was at my former employer. SIs asked shotgun questions and provided no debug data, then bitched when we didn't answer the question "correctly". The good ones usually got their answers in half a day because they gave a debug dump, clear description of the problem & what was expected, and dead perfect reproduction steps.
MBA / EE is a killer combo.
JD / EE is another one.
I have a 36 hour cron job rsync that runs on my critical work.
Everything else is weekly.
Plus a pair of local drives that take turns being on the system every other week.
define "Survive"...
In that scenario I would lose 1 week's worth of work at a maximum.
On critical projects I would lose ~24-48 hours worth of work.
Of course I find the situation to be highly unlikely... so am willing to accept that risk level.