New Rochen.

Artashes

Artashes

Administrator
Staff member
I have just noticed that Rochen hosting (http://www.rochenhost.com) has just adapted a new look and new service.

The web site looks pretty, but could someone please explain to me how exactly RAID 5 and hot-swappable SCSI hard disks would guarantee for more redundancy?
http://www.rochenhost.com/premium-hosting/

This means that in the event of a hard disk failure you should experience no system downtime or loss of data.
Click to expand...
Does system downtime basically mean the same as web site downtime?

Thank you.
 
Does system downtime basically mean the same as web site downtime?
Click to expand...
I'm no expert at this, but in this context I believe it does.

One hard disk goes down, yet with Raid you still have all the data and the server continues to function relatively normally (there might be a performance drop). You can then change the faulty hot-swappable hard disk while the system is running and rebuild the array on the new disk. Bottom line is that you have zero downime caused by the disk failure.

I just hope I got things right, because I as said, I'm not an expert at this.
 
no one can guarantee 100% uptime. even warner bros, ms and other companies cant. Software needs to be updated and system has to restart.

further more, if 1 HD goes, the 2nd one takes over and so on ... but you can advertise it, and you might never know if its real.
 
Well, the way I see it, you can guarantee 100% uptime if the solution can handle it (suppose you have a farm of servers on various networks worldwide - you ought to stay up in case of one network failure). Not that Rochen advertises 100%, but just a thought - it all comes down to how much money you are willing to pay to stay online.

I guess hard disk failure prevention is a great idea, however, seems like it won't protect from network downtime, which is often a more casual reason for web sites to go offline.
 
Yeah, I've learned a lil from this thread and now I know why its always 99.9% uptime, not 100%.
 
This means that in the event of a hard disk failure you should experience no system downtime or loss of data.
Click to expand...
That all they're saying. That if a hard disk fails, that will not result in downtime. However, what is going to happen if two HDDs fail at the same time? ;)

Well, the way I see it, you can guarantee 100% uptime if the solution can handle it
Click to expand...
Unfortunately there is no solution to achieve 100% uptime indefinitely.

I guess hard disk failure prevention is a great idea, however, seems like it won't protect from network downtime, which is often a more casual reason for web sites to go offline.
Click to expand...
A hard disk failure can lead to hours of downtime. It is also an event that vitually all hosts who have been in business for a while and have a certain size, had to deal with. It is no joy.
 
Last edited:
ldcdc said:
I'm no expert at this, but in this context I believe it does.

One hard disk goes down, yet with Raid you still have all the data and the server continues to function relatively normally (there might be a performance drop). You can then change the faulty hot-swappable hard disk while the system is running and rebuild the array on the new disk. Bottom line is that you have zero downime caused by the disk failure.

I just hope I got things right, because I as said, I'm not an expert at this.
Click to expand...

Dan,

You got it right. When you use a SCSI RAID controller and implement RAID5 you need a minimum of three drives. With RAID5 the data and parity are striped across the drives. This way if any one drive in the array fails the controller can "rebuild" the missing data from the parity information. It is a very reliable method. Most all of the servers I set up for business use RAID5 with a hot spare. This way if a drive fails the controller brings the hot spare online and rebuilds the "missing" information on it. There is a brief period of reduced performance while the drive rebuilds, but for the most part users never even notice. If two or more drives fail at the same time then RAID5 will fail, but that is very, very unusual. In the few cases I have run across where data integrity and availability required it I have used RAID51 (5+1). An eight drive RAID51 array can handle 3-5 drive failures before losing anything.
Modern SCSI drives are remarkably reliable and actually do an amazing job of predicting failure. I've surprised many a client by showing up with a replacement drive and telling them that the server has predicted a failure. I can't speak for other vendors but I know HP/Compaq will ship a replacement drive under warranty on the basis of a failure prediction.
 
Last edited:
AFAIK they had issues too. I don't know if they were hardware related though.

Here's an interesting article about a big company that absolutely needs a stable system and certainly has the money to pay for it:

Last year Visa pumped $2.3 trillion through its 15-million-kilometer matrix of fiber lines. In five years it has suffered only eight minutes of downtime, better than almost any other system on the planet.
Click to expand...

http://www.forbes.com/global/2002/0916/038.html

Edit: I just read your post Thomas. It was a very interesting read. I'm also glad that I didn't say something stupid in my previous post. :)
 
Last edited:
That Forbes article on VISA was an interesting read. 4000 transactions per second is an amazing thing to accomplish when one considers all that has to occur for a transaction.
 
Thanks Everyone!

Hey Everyone,

Thanks for all the comments :)

I think it is also important to point out that while RAID can certainly add redundancy in terms of service uptime in the case of a drive failure it is still very important that regular backups are captured too. RAID can not protect against things such as file corruption because the errors will be written to all of the disks in the array. Bottom line is, RAID is not a solution on its own.

Taking this into account, with our premium web hosting plans we make nightly backups to a another drive (which is not part of the RAID array) and capture off-site (not just off-server) backups twice-weekly. Between the RAID 5 disk arrays, nightly backups and twice-weekly off-site backups our customers' data should be pretty safe. Of course, we still stress the importance of customers making their own backups too.

For this particular product we are using Dell PowerEdge 1800 Series servers.

We are planning to start an aggressive marketing campaign for our new plans soon. However, we have had a large flood of upgrade requests from existing clients wanting to use the new premium service. As such, we are meeting this demand from our existing customers before we start marketing the premium service to protective new clients. Customer service is one of the most important things here, so we don't want to bite off more than we can chew.

I had an extensive chat with Artashes a few weeks ago, so you will probably see some Rochen banners ticking over here at Hosting Discussion amongst other places :)

- Chris
 
You must log in or register to reply here.

Supporters

Back
Top