Comment OS that uses text on a black screen (Score 1) 81
Isn't [Linux] the hacking OS that uses text on a black screen?
I'm sorry. You must have Linux confused with Windows Server Core.
Isn't [Linux] the hacking OS that uses text on a black screen?
I'm sorry. You must have Linux confused with Windows Server Core.
RAID itself isn't a backup. However, using multiple disks with btrfs or ZFS is approaching it. They address the case of most pilot error failures as long as you actually make snapshots (no more rm -rf disasters) that prevented RAID from being the answer. You are still left with a few disastrous failure modes like the power supply blowing up just wrong and putting AC line voltage on both drives or a fire, but it is approaching.
You can do cross backups between two machines to eliminate the power supply failure mode. For most home users, regular off-site backup of that much data simply won't happen.
Marketing weasels.
That's why I celebrate the arrival of the 6 TB drives. They really brought the price of the 4's down.
The truth is they're all at the bottom of the barrel. Nothing is worse than X except for Y. Now Y is really crap. Only Z is worse than Y. Never Use Z. The only thing worse than Z is X. For every major brand out there you will find glowing reviews and horrific failure reports in about equal amounts. They all have good runs and bad runs. Occasionally they will have a particular model that is nothing but fail.
That's why for larger systems you should use multiply redundant arrays. For example, RAID6 or 3 way mirroring. That way you can cover the increasingly probable case of losing a disk while the re-construction is in progress. It also becomes increasingly important to use drives from different batches and preferably different ages.
It's also helpful to have spares on-hand. I would like to see a concept of warm spares where the designated spares do not get powered except for periodic testing and when actually required so the system doesn't have to wait for you to notice the failure and put hands on it to begin reconstruction AND it doesn't put excess wear and tear on the spare drive.
There are a number of workloads where caching is not so useful. For example, video conversion or 'big data' analysis where you are streaming the inputs. At that point, an SSD is more of an intermediate buffer than it is a cache (so only helpful for writing). If your use pattern streams more data out than the size of the SSD, then it's only getting in the way.
In a file server, unless you are using multiple gigE or faster interfaces, having plenty of RAM will make a much bigger difference than SSDs will.
Not if China blocks the VPN's handshake, as it has been seen to do.
There is very few censorship agreements in the USA. Most of the so called hidden agreements are just the morals of the particular publisher.
And some are laws that allow a publisher to inflict its will on third parties, such as intellectual protectionism/imaginary property laws.
Except they have a capitalist vision of google having to pay to have people access gmail.
wat. Of course Google pays to have people access gmail - servers, racks, drives, power, transit, staff, real estate all cost quite a bit.
Chrisq's claim, as I understand it, is that Google would have to negotiate transit with each individual last-mile ISP to make Gmail available to the ISP's customers, or at least available at more than 1999 DSL speeds.
Obviously serious problems only could occure if software was not patched.
If one claimes even patched software will cause serious problems (your preparations) he is an idiot.
I think the people predicting Y2K doom were under the assumption that someone somewhere might miss a very important patch.
whatever happened to: If you don't like the way a company does business, don't do business with them?
Monopolies happened. A lot of times, there are no compatible substitutes for a particular company's products or services despite the company's poor information security practices.
If they're storing the passwords in clear text, that's not good. However, they could be assigning random passwords and only storing the hash after they send it via email to the user. There's just not enough information to say.
Agreed that security questions in addition to the usual click lost password and they send you a unique URL to navigate to is a good idea and considerably improves the security of password recovery as long as the answers to the security questions aren't easy to determine from looking through the users email box.
There are also situations where systems integration must be performed without single sign-on.
And even with single sign-on, there are situations where a system participating in single sign-on needs to store a "client key" and "client secret" for something like OAuth.
So to fix this we added the "get connected" feature. Basically it's a page after the initial login where people can open a session to all their social networks and provide all their frequent email addresses. This way they can login with any of these. This helped a lot.
The Stack Exchange network has a similar feature. Each user can associate a Facebook account, an e-mail address and password, and multiple OpenID identifiers (Google, AOL, Ubuntu, etc.) to his Stack Exchange user account. The one thing I'm surprised they don't support is Twitter login.
Intel CPUs are not defective, they just act that way. -- Henry Spencer