when it comes to providing redundancy for hard drives, the first thing you should think of is raid
redundant array of independent devices,
or you could hear independent disks.
Let's find another way.
The idea is, there's redundant array.
We have multiple disks acting as a single logical disk,
and it's usually for the purpose of redundancy.
That being said, the very first raid we look at Raid zero is not redundant.
If I ruled the world, I would call it a day.
It's an array of independent disks, but it's not redundant
without redundancy. What that means is, if one of your disks fails, you run the possibility of losing all your data
when Raid zero does is it takes two physical disks and uses a feature called Disk Striping.
If there's 24 kilobytes worth of data to be written to the desk, 12 kilobytes gets written to disk one and simultaneously. The other 12 kilobytes is written to Disk two.
You get a simultaneous right instead of a sequential right.
It saves time. It's a speed improvement.
Not only is it faster for writing, but it's also faster for reading because his data is being located on drive one or is being read on Drive one. It's being located on drive to
You get a performance boost, but because of the fact that it's not redundant is not really a great choice for most servers.
This would be good, however, on media streaming servers where we get that performance. But we don't really have the need for fault tolerance
if high availability is my need, Raid one.
What Raid one is is it's disk mirroring.
I have two discs, each one an exact replica of the other.
This provides really high availability, because if there is a failure of one of the disks, it's very easy and very quick to transfer over to the other.
You hardly have any down time at all.
Drawback. There is is I'll always lose half of my disk space.
By that, I mean, if I go out and buy 25 terabyte drives,
I have 10 terabytes of space.
Five terabytes is for redundancy, so I don't get to use that person.
That's kind of the one of the drawbacks to this marrying but really good redundancy.
Raid five disk striping with parity,
this takes a minimum of three physical disks
I have the striping, which gives me the speed improvement but also on each disk. Contains parody for another disk.
Ultimately of disco unveils parody on disc two can be used to rebuild the disk that's failed.
That being said, this is kind of a dicey process, and sometimes rebuilding the disk can actually cause data loss or disk failure.
Raid six is kind of coming around to replace Raid five.
Here we need for physical disks
by using for physical disks. Two discs are for fault tolerance so that as many as two drives can fail.
That's a little bit better of the redundancy
ray 10 or sometimes referred to as one plus zero. We said this striving
the stripes that isn't fault tolerant.
What if I'm near that shapes that to another set of disks.
That's what read tennis
that requires four physical disks as well.
Of course, half my Swiss would be used for fault. Tolerance
rate is always discussed on the exam.
Expect to get questions.
Expect to get a couple of performance based questions
that shall know thy raid.
Another hot topic is going to be backups because even though we have rate in place file corruption, malware can infect the entire raid array.
Ray doesn't really protect our data that way. We needed to be protected, so we back up our data.
There are multiple types of backups we can use.
There is a full and incremental a differential and a copy.
The full is the easiest to understand because a full backup backs everything.
Incremental backup backs up what's changed since the last backup of any kind.
We do a full backup on Sunday night.
We do an incremental on Monday. That backs up was changed in Sunday.
We do. Another incremental on Tuesday and backs up was changed since Monday.
An incremental Wednesday backs up, which changed since Tuesday.
This takes less time and back up.
However, in the event that we need to restore, we have to restore the full backup in each day's incremental.
That can take a little longer.
A differential backup backs up, which changed since the last full backup.
We do a full backup on Sunday.
Monday's incremental and backs up was changed since Tuesday,
Tuesday's incremental, which changed since Sunday,
Wednesday's incremental, which changed in Sunday.
This is going to give us a quicker restore because we have to restore the full backup and the most recent differential.
All this has to do in Windows environments with a little bit called the archive bit.
The archive bit is just a flag that pops up and says have changed.
I need to be backed up.
When you do a full backup, all the bids get reset.
That's just a way of acknowledging I've backed up everything.
Everything that needs to be backed up is backed up
on Monday as files start to change, those flags pop up.
We do an incremental Monday night, and that backs up everything for the day, and it clears the bits and signals that's taken care of.
Tuesday's new files are modified. The five pops back up
Tuesday's incremental backs up, which changed on Tuesday and clears the bits.
The differential backup is different from the other two because the differential bit backup does not clear the bit.
Which is exactly why Monday's files changed.
It backs up everything that's changed on Sunday. The bit isn't cleared.
This files pop up Tuesday. You have the things from Monday with flag set and you have the things from Tuesday.
We back up everything since the last full backup.
with virtual machines. You can copy specific files or you could do a full backup.
This just does not reset the archive bid.
Also with virtualization, we now think about just reverting to snapshots.
Whatever a strategy is, we need to make sure that we have fault tolerance for our data as well as our hard drives.