Posted on 09/16/2016 10:13:55 PM PDT by djf
Hi everyone, I have a question about SAN disks, which is a standard electronic disk. I have two of them, one is 128GB and the other is 240GB.
Many sites report that electronic media do have a limited number of writes/reads before they quit.
I have never been able to push one of my drives to fail, and don't really want to destroy something on purpose. But I am intensely curious what happens when they do fail.
Is it smart enough to report to the OS (operating system) that there has been a track failure? Will the OS keep operating, and a re-write attempted? Does anyone at all have experience with this?
Thanks in advance!
I know that with Linux you can use the SMART interface if the drive is capable of it. I check mine periodically to see if I’m showing persistent read or write failures. Most of the time if a sector goes bad, the drive will automatically mark the sector as bad and move the data elsewhere. However, when you see lots of these rewrites, it is time to replace the drive. It is always better to do that ahead of time in a planned manner rather than an ‘oh shit’ moment which tends to be at the worst possible time. I’m sure there is a similar utility with windows but I would not have a clue what it is.
My career is CNC repair. Computer Numerical Control, in other words, computer controlled machinery. All of them have disk drives for data storage. When the disk fails it is usually the motor that spins the disk that stops working. At that time, one is unable to read or write to that disk.
True........
bookmark for later
So SSDs (solid states) as opposed to magnetics. Burn a little bit of material away everytime data is written. Each cell of data has a certain number of times it can be written to until there is no material left. In order to protect against the same cell being written to all the time (lets say you keep resaving the same file). A little controller chip on SSDs spreads out the writes to different places on the SSD all the time.. This is why there is no reason to “defrag” an SSD.. They are by design fragmented.
Anyhow.. If you manage to burn out a cell, there is what is called over-provisioning. In other words, your 240 gig SSD actually has 260 gig.. Its just the addition is reserved for when cells fail or burn out, they are transparently re-assigned to cover for the failed cell.
S.M.A.R.T as some people have mentioned will track that statistic.. When cells burn out and the over-provisioning is being used to cover for the bad cell, SMART will notify the operating system that the drive is beginning to fail and you should move your data.
Over provisioning is one of the big differences between consumer and enterprise grade SSDs. Enterprise grade SSDs will have huge amounts of provisioning.. They also use cells that can handle way more write/erase cycles than consumer brands.. For instance, the difference between the Samsung 850 EVO and 850 PRO is really just how much over provisioning is available.
Ain't quantum mechanics the cat's meow?
Post of the thread!
You’re funny!
Bump
I have a copy of my data backed up on two external hard drives . . . yes, I had one fail and lost everything - so now I use a system of replication. If one fails, chances are the other one isn’t going to fail at the same time - just out the money for a Seagate, unless you buy it at Costco where you can just go and trade it for another one and then reback-up the data!
Yep.
If one of my drives takes a shit, I’m still good.
Backed up to two other computers and a couple thumb drives as well...
And now the cloud.
and that’s not all. I have a copy imaged in VM, so all I have to do is load that image on a computer and voila!
Yup. I used to do a lot of application programming on dumb terminals tied to mainframes. They were real stupid, in that you had to specifically save your keyed-in data periodically with a command. If you took a minute (or seconds) more than you should have, the terminal would lose its connection or something else would happen to the mainframe, and your most critical data was gone forever. The terminal would often sense when you were about to do a save, and lose the data a second prior to that. Now I use software that periodically saves my data, my data is automatically backed up every hour, and my hard drives are in RAID arrays on UPS back up battery systems. The ghosts in the machine look for opportunities to screw with you unless you beat them at the game.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.