Some anectodal evidence: Five years ago 2TB was the highest capacity drives. I bought 3 pieces of 1.5TB drives, so not the biggest, but next to it. They were the cheapest drives of the my usual manufacturer. One almost immediately failed. The replacement drive also failed within a year (different type, but same manufacturer). A third drive is still working but after a power loss a year ago quite a few bad sectors appeared on it, some data was lost. All in all only one from the four had no issues within 5 years.

On the other hand, I had no problems with many other drives over the years from the same manufacturer, which were also SATA drives, but medium capacity, relatively more expensive models with longer guarantee.

For me the lesson is that I should not buy the highest capacity drives of their generation, because the technology may not be mature enough at that point.

It is funny, that on one hand you (or who knows, maybe another anonymous coward) use the cheapest consumer HDD prices you found at the cheapest places in your examples, and on the other hand you continuously use extra or not even existing future hardware when you talk about features.

There are hard disk cartridges, specifically made for backup, with better reliability than a standalone HDD, but in this case they are not that cheap either, and they are still less compatible with backup software: http://en.wikipedia.org/wiki/RDX_Technology

The required backup capacity depends on how quickly your data changes. If you have a quickly changing 6 TB, then for a reasonable backup you will need about 60-120 TB of backup within a year. In our case, reasonable means daily backups preserved for one week and monthly backups preserved for one year. (And yearly backups preserved forever, but I do not count them.) It did happen to us that a coding error rendered important data useless, in such a tricky way that we had not noticed it for a year.

If you have the ability to cheaply and conveniently backup a large amount of data, then you start to backup things which would not even have occurred to you previously. For example it could have been useful if we had HTTP logs from several months earlier when we talked to the police about an attack on our system. And according to my - admittedly limited - experience, tape is the medium which is cheap and convenient.

But it also takes only 5 seconds to delete all your data.

Sequential access speed is only relevant if you backup huge non-fragmented files or entire raw partitions, and nothing else.

Yes, they are surprisingly fast. The maximum speed of a current Tandberg LTO-6 drive is 160 megabytes/s if the data is uncompressable. With the usual compressible data it can be about 320 megabytes/s (officially 400).

These drives can even be too fast. The drives do speed matching, but they have a minimum speed, below that they start shoe-shining. One reason I have chosen an older generation, LTO-3 tape drive, instead of the current generation, because I cannot easily feed an LTO-6 with at least 60 MB/s, which is the minimum speed of the drive. Considering compression, that is about 120 MB/s, which saturates a 1Gb network.

Capacitors make possible to cache already written and synced data on the drive. For example, you write many updates to a single file, like in case of the MySQL replication status file. If you cannot lost even only a few writes, then you must flush all these writes to the disk platter / flash memory. This is of course really slows down things. And quite unnecessarily, because everything which is written out, will be overwritten within a few milliseconds.

If you have capacitors on the drive, these small writes never reach the flash memory (except on system shutdown), because the drive can safely store them in memory. If there is a power loss, or other problem, the capacitors provide enough energy to write out the content of the cache to the flash memory.

Capacitors are the smaller equivalents of the battery backup units in RAID controllers.

I also considered 840 Pro, because I assumed that "Pro" means that it has capacitors, but no, it has not. Absolute performance is misleading alone. It must be considered together with reliability and consistency of performance. These three often represents trade-offs. It is easy to create a drive which is very good in random IO: use a large write cache on the drive without capacitors, and lie to the OS about sync. Manufacturers have done this previously, maybe they do this today, they do not talk about the internals of the drives, I do not trust them. I ended up with Intel DC S3500, contrary to the fact that I am not a fan of Intel. It is a server drive, so I hope it does not lie to the OS, and the price is not much higher, if at all. It is not optimized for the "desktop", but it has a consistent performance. I haven't even checked absolute performance, but I am sure it will be fast enough for me (because a hard disk was also enough).

The firmware is not interested in filesystem pointers. It does know, however, if a block has never been written.

TRIM is essential for maintaining SSD performance.

This is not so simple.

The original TRIM command is non-queued. It can kill drive performance on servers, so enterprise drives are designed to work well without TRIM. If you want better, and more importantly consistent performance then you should overprovision the drive. Overprovisioning means that you do not partition 20-40% of a new drive (or a used drive, after a secure erase). Those blocks will never be used, therefore the drive always have plenty of free space, so there is no need for trim.

Queued TRIM command appeared only in the SATA 3.1 specification, so only new drives support it.

We are talking about a short, almost personal comment on the developer's mailing list of Ubuntu:

i personally wouldn't do online banking with it ;)

Compare this with the Slashdot article title:

Canonical Developer Warns About Banking With Linux Mint

Whether he is technically right, or not, I find it disgusting that such a side note becomes news on Slashdot.

By the way, the subject was another new distribution based on Ubuntu, similar to Mint, therefore the Ubuntu developer actually encouraged an Ubuntu derivative.

My mistake, I wrote "script", but it is not, it is a job configuration file. The exec line defines the executable and its arguments. The example is a Minecraft server, which is written in Java, so the executable is java in our case.

The Upstart reference documentation is http://upstart.ubuntu.com/cookbook. If you want man, then man 5 init describes the job configuration format.

The Upstart documentation was indeed poor a few years ago, but now it is complete and easy to read.