Oh boy... the next fews days are going to be painful

[PCBrandon]

Hi all,

Basically for three years we've been putting band aids on our main server on my campus. The server handles email for over a thousand people, internet access, a website, Norton corporate, and a whole lot more.

Well, we hit a snag late last night. We ran out of band aids, and the server just doesn't like us anymore. The RAID driver crashed and burned and brought Windows with it. We were able to stick in an emergency boot disk to pull off the data in four of the six SCSI hard drives and got with Dell to send us a new motherboard tomorrow morning. The IT admin there is hoping this is going to solve all his problems once they replace the board, but I know it won't. We are going to have to start fresh, which it really needs it anyway. We've already got all the data backed up, so after the motherboard is replaced we will go on with the software.

We have to install:
Windows 2000 Server (We don't have a 2003 license for that server, we have it on another server at a different campus)
Critical Updates for Win 2000
Exchange 2000
ISA
ASSP (A spam filter for email)
IIS
Norton Corporate
Windows Update Services (for regulating what critical updates are allowed to be installed on the computers in the domain)


Anyone have any advice on what order we should install it in? I have very little experience with this server stuff, so I'll basically be going by a book for majority of the time.

[mairving]

The biggest problem is having too many services on one machine. I would say that you really need a minumum of 3 servers. ISA on one, Exchange on another, IIS on another. Personal preference is also Trend Micro Client/Messaging Suite for A/V and GFI MailEssentials for spam filtering. It does sound like fun though.

I will also have fun this weekend. I will be upgrading our SQL Server program that nearly everything runs off of at work. I will have to do 3 different version upgrades on it. On a test server, the first upgrade took about 10 hours. The other 2 were about an hour a piece. Shouldn't take near that long on the production server. Still should be a fun weekend.

[PCBrandon]

That is exactly what I told him. We have a bunch of servers that are doing pretty much nothing.

I'll give them names for the benefit of this post:

mailserve
iserve2
bob


Mailserve is the current server that handles everything. I want it to only handle email and have iserve2 do all the routing with ISA. However, the IT admin says they've tried before and can't get external email to route through iserve2 to mailserve. I think that is should work, why wouldn't it?

Mailserve would be the domain controller (MCKEELACADEMY.COM) as well as the email. The local website would be on www2.mckeelacademy.com on mailserve. How am I going to set it up so it works like that again? It also had OWA (www2.mckeelacademy.com/exchange). www.mckeelacademy.com is on a seperate web server that we don't control.

If I were to move the website onto Bob, how could I get OWA to point to the mailserve for email? Also, if I were to put ISA on iserve2, how would I get it to route external emails to mailserve?


Wow I'm going to be learning a lot in the next few days.

[PCBrandon]

Today wasn't too bad. It was looking horrible in the morning, but by the afternoon things were looking up.

We arrived on campus at around 7:00 AM to find the new motherboard, floppy disk drive, and floppy cable already there (apparently we haven't been able to boot to floppy for years and it turned out to be the controller on the motherboard). The technician arrive around 8:00 AM to replaced the hardware and was out before 9:30 AM. So, we hopped on and did a BIOS flash and updated the PERC, EMS and Backplane firmware. After we rebooted and was about to configure the new array containers one of the SCSI drives started flashing amber lights. Shortly after, a 2nd drive started to flash amber lights. So, we went into the RAID Utility and did Drive Diagnostics on each of the six SCSI drives, it found all the sectors on each drive, including the ones showing amber lights to be good. So, we stuck the drives in the final configuration:

00 - 18 Gig 10k RPM
01 - 18 Gig 10k RPM
02 - 36 Gig 10k RPM
03 - 36 Gig 10k RPM
04 - 36 Gig 10k RPM
05 - 36 Gig 10k RPM

We put 00 and 01 on a RAID 1 container for the OS and software. That way, if something like this should happen again we don't have to rebuild the whole thing. Drives 02, 03, 04 and 05 are on a RAID 5 container for all the Data for exchange, ISA, IIS, and more.

The utility had to scrub the drives after we made the container. The OS container (00 and 01) finished but the Data container (02 - 05) was still at around 40% when we left, seeing how it had 101 Gigs to go through.

Once we arrive there on Monday we'll start with the OS and other software. I am kind of worried that all of the software will take up more than 16 Gig (since the two drives are mirrored we only have 16 usable Gigs). We are putting SOME stuff onto another server Bob... mainly the staff folders where they can upload for "safe keeping". Also, we are thinking about taking away middle school email accounts, and possible high school as well because if they do use their accounts, they use them like an IM service.

Email 1: "hi"
Email 2: "wazzup?"
Email 3: "nothin much"
Email 4: "you?"
Email 5: "this class boring man"

and so much more...