Reliable Web Server

[s0laris2]

I "HAD" a Raid 5, 4 drive IDE system with Maxtor Drives.

Last night 2 drives died instantly.. I lost EVERYTHING as it was 200 gigs of data and that is pretty hard to back up. I just started out in the biz and have to start all over again. I have all my content still at home, but I have to re make all the sales pages, etc again :-((

My Question is what is a good reliable webserver I can buy that us Rackmount and Runs RedHat 7.2. I have no problems setting up the Redhat and harwarem, etc.. I just want something that will give me piece of mind so I do not have to go through all this again :-(((

I was looking at a Dell WebServer... Any suggestions?

My Doimain will not come up as the server is shutoff since it wont boot anymore.

[-=HUNGRYMAN=-]

If you want piece of mind ... DELL Poweredge is the ONLY way to go ...

[Pornwolf]

That's too much work to start over man. Have your host do a disk recovery job on those drives. It might cost a lil extra but it would be worth it.

[s0laris2]

Its my own server. I built it and shipped it. Yes I can probably get a new circuit board on 1 of the hd's, but the second one is KNOCKING.. so my platter is compromised

[-=HUNGRYMAN=-]

d00d !!! Youre getting a Dell !!!

[edmo]

ibm xservers are sweet.

[RK]

Originally posted by s0laris2
I "HAD" a Raid 5, 4 drive IDE system with Maxtor Drives.

Last night 2 drives died instantly.. I lost EVERYTHING as it was 200 gigs of data and that is pretty hard to back up.

That is very unusual (and highly improbable). Is it possible that someone got into your server (physically or over the network) and did it?

200 GB of data? That is a lot for web sites. Maybe a tape back-up or DVD-RW system is what you need? and backup only things you need (no huge log files etc).

"Yes I can probably get a new circuit board on 1 of the hd's, but the second one is KNOCKING.. so my platter is compromised"

So it's not that bad. If you have 3 out of 4 hard drives from a RAID 5 array, it should work and you will get all your data back.

As for recommendations, almost all our boxes are custom built but from my experience (for IDE drives) IBM is great.

[Juge]

You had RAID and 2 drives crashed at the SAME TIME!?!

Dude, buy a lottery ticket.

[whoreans]

maxtor drive = crap

[ZakT4PH]

dell poweredges suck, stile just got two 2650's and first off the BGE drivers for the gige ethernet didn't work correctly massive packet time outs. two problems. The main was delt with the new PCI-X busmaster? system that they use. Even with standard intel pro cards (fxp drivers) they had packet loss. it turns out that you have to apply some custom patches to the freebsd kernel and change some other modifications in the kernel source so far we have been testing them for almost 24 hours and hope to see 0 packet loss and the testing finished. what a waste of time for scsi drives seagate and fujitsu make good drives but IDE seagate is the best

[vending_machine]

Host somewhere where you can back up your data, 200 gigs is a lot, but not unmanagable.

If you run with IDE, expect drive failures. We have a backup server running RAID on IDE drives and drives die about once/month.

Instead of going with RAID 5 on SCSI, I'd just have two servers and mirror them, you wont get better redundancy than that locally.

[-=HUNGRYMAN=-]

Originally posted by ZakT4PH
dell poweredges suck...

alright .... tell that to some of the biggest datacentres in the country ...

ANYBODY could get a bum piece of hardware, at any given time ... but you tell me DELL did not take immediate action to rectify the situation ... ???

The didn't become what thy are today by sucking .... or they'd be called MADONNA Computers !

[DrGuile]

Originally posted by RK
As for recommendations, almost all our boxes are custom built but from my experience (for IDE drives) IBM is great.

I would stay FAR FAR away from IBM IDE drives!

they're deskstar serie had faults in them. they would die like crazy. There's a class action lawsuit against them.

[Goonx]

I B M

[RK]

Originally posted by DrGuile
I would stay FAR FAR away from IBM IDE drives!

they're deskstar serie had faults in them. they would die like crazy. There's a class action lawsuit against them.

Yes, I'm fully aware of this. That is why I said "from my experience." I saw many less failures of IBM drives than Western Digital and many others and IBM drives are fast.

[Brad Mitchell]

The stupid comments I see on GFY never cease to amaze me.

[RK]

Originally posted by SinEmpire
The stupid comments I see on GFY never cease to amaze me.

Can you be more specific?
Personally, I would appreciate your input.

[eru]

Originally posted by pornopete
It doesn't matter what the sticker on the outside says.

If you want to setup your HD's in a raid5 configuration then you should be using SCSI.

It will cost you a lot more.

Oh and redhat is crap. See if your host will install FreeBSD for you.

Long live SCSI -- I use it on my home computer (15000RPM baby!). Relability, stability, and the catch? Money.

[Jamie]

Originally posted by -=HUNGRYMAN=-
If you want piece of mind ... DELL Poweredge is the ONLY way to go ...

Nothing better than the DELL's

[Brad Mitchell]

Originally posted by RK


Can you be more specific?
Personally, I would appreciate your input.

There's lots about servers that would be worth talking about... The only way to determine an optimal server setup is to first identify what it will be doing (scripts, html, etc) and next how much utilization (bandwidth).

Here's a few miscellaneous facts..

15k RPM SCSI drives are cool but they're way fucking hot. There's no way I'd put those in a server because I think one would be unlikely to appreciate the difference since things are generally slowed by the system bus or other components. Either way, you had better install additional cooling.

IBM drives suck? Oh come on, give me a fucking break. They are excellent drives. All manufacturers have a small failure rate - most likely you ended up with shit-ass bad luck. When purchasing drives, IDE or not, look at the specs and the MTBF (mean time before failure). That will give you some idea of drive quality. Low quality drives may be 100,000 hours while higher quality manufacture will generally result in 3-5x that.

Raid 5 with 4 drives? That's interesting. Typically people do them with 3 drives or 5 drives (I'm not familiar with higher numbers as my largest raid 5 had 5 drives). Perhaps someone that is more of an expert can tell me if there is something odd about that.

It sucks that two drives died but I'm wondering if it was due to human error. That would otherwise be a real statistical anomolie. One thing to consider here is that even with a RAID system having a backup to tape is a good idea. Remember that despite the fact a RAID configuration will protect you from drive failure it will NOT protect you from human error, viruses or a hack.

FreeBSD is probably superior but hundreds of thousands of servers operate successfully on Red Hat so to say it's crap is nonsense.

More advice? Leave building your server up to an expert. Why? Well, it's broken now isn't it? lol

Zak - Sounds like you had bad luck... or a bad install... and that sucks. However, Dell does not. It's nonsense to say they are shitty computers when everyone from NASA to LikeWhoa uses them. There's statistical error in the manufacturing of everything. Beyond that, if you're having configuration problems then welcome to the world of setting up servers - doesn't matter who the manufacturer is or what the OS is. Shit happens.

I'm a total fucking geek. I was probably the only 14 year old with every issue of Computer Shopper in 1989. The reality is that there is good and bad in everything. In the world of computers failure is statistical and installation problems are commonplace. Is there hardware I don't like? Sure, tons of it.... but that's another thread.

Cheerio!

Brad

[gothweb]

Have to say it... otherwise my head will explode.

http://www.apple.com/xserve/