Hardware / Backup Procedures

[logic404]

Hi,

I realise that given the problems of the last two days, this is probably a singuarly BAD time to ask this! So, I'd like to preface it by saying that although I had a small panic attack this morning when my site was down again, I have enough trouble keeping the SOHO network running and have nothing but respect and admiration for the techs who have busted a gut to keep us up and informed as best they can.

My question...

A prospective client asked me something along these lines the other day, "Why would I want to host with you when I can put my site on my home computer over my cable connection?". I almost laughed at him, and then began to explain that what you are paying for is people that look after your server, a commerical level connection etc. So then he starts asking for specifics - what happens if my server goes down? what happens if the harddrive failes? Etc.

I'd like some correction and perhaps injection of information on the following which I told my prospect:

1. That Vortech have a backup connection to the 'net - this was a wild ass guess, but I assumed that if ISP A who supplies your connection loses the plot, then you switch to ISP B - total downtime maybe 2 minutes?

2. That Vortech have backup machines to hot swap in the event of hardware failure. I've since read that you have about two, and that this solution (reasonably obviously) only works where the DISK is not the damaged component. That means, tho, that say the MOBO dies or something, you can just swap in the new machine with the old disk, no harm done?

3. Backups. I didn't make any assertion here because I have no clue. Do you do any backups of our data on your servers? If NT28 (the machine I'm on) died tomorrow (as in the disk), what would happen?

Thanks in advance guys. I just want to understand what all the procedures are in order that I set the expection level with my clients correctly.

[admin]

1. That Vortech have a backup connection to the 'net - this was a wild ass guess, but I assumed that if ISP A who supplies your connection loses the plot, then you switch to ISP B - total downtime maybe 2 minutes?

A: The main core routers are juniper M5's and yes they are the ones going right to the providers. There is 2 of these. What has been going on with the router issue has been the 2nd router for the network. These are the ones being replace tomorrow AM. There will now be one with failover with a spare on hand as well and maybe a few others just to be sure.


2. That Vortech have backup machines to hot swap in the event of hardware failure. I've since read that you have about two, and that this solution (reasonably obviously) only works where the DISK is not the damaged component. That means, tho, that say the MOBO dies or something, you can just swap in the new machine with the old disk, no harm done?

A: 95% of the time we have the same hardware on hand thats in every system. If not we can get it working farly quick in case of MB, HD, Powersupply and such.

3. Backups. I didn't make any assertion here because I have no clue. Do you do any backups of our data on your servers? If NT28 (the machine I'm on) died tomorrow (as in the disk), what would happen?

A: Yes we have 2 packup systems one 600GB and one 1000GB backup system both raid5.

[logic404]

Thanks Brad. The only bit I didn't quite understand was your answer to my question about the backups? What does it mean:

A: Yes we have 2 packup systems one 600GB and one 1000GB backup system both raid5.

As in, the data for our websites and stuff is backed up here? How often? I'm a tad 'fused...

We take daily backups of web and database content. These backups exist for the express purpose of data recovery in the event of catastrophic media failure on the servers. Users are still advised to keep their own backups since ours are designed for internal use and we'd have to bill you if you needed to recover data if you or your customer are responsible for the loss.

[logic404]

Not a prob - I advise everyone to keep backups anyway, but your answer was what I was looking for - that in the event that the data loss is quote unquote "your fault" (e.g some hardware failure, which is probably not really your fault, but it ain't ours either!) that the last backup you have would be restored, and it's up to the client to put back any change they made in the last 24 hours from their own backups.

Thanks muchly guys!!