• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Ordered a file server - Did I do the right thing?

S

Sasiki

2[H]4U
Joined
Apr 10, 2005
Messages
2,818
Norco 4020 case and 26" Norco rails - http://www.newegg.com/Product/Product.aspx?Item=N82E16811219021&Tpk=norco 4020
Corsair 850TX power supply - http://www.newegg.com/Product/Product.aspx?Item=N82E16817139009&Tpk=850tx
Asus P5BV-E server motherboard - http://www.newegg.com/Product/Product.aspx?Item=N82E16813131252&Tpk=p5bv-e
Intel Xeon x3220 2.4GHz CPU - http://www.newegg.com/Product/Product.aspx?Item=N82E16819117127&Tpk=x3220
WD 500GB RE3 HDD's x 4 - http://www.newegg.com/Product/Product.aspx?Item=N82E16822136294
Crucial 2x2GB DDR2 EEC ram (2 kits) - http://www.newegg.com/Product/Product.aspx?Item=N82E16820148115&Tpk=ct2kit25672aa667

Total cost charged to card - $1425

I will be using this server as a corporate file server for our main office, 35 users (only about 10 heavy users). All user Document folders and Outlook PST's will reside on the server. The hard drives will be configured in a RAID10 array. If VMware will work with the Intel ICH7r controller, I will do that and then Linux within the VM. I did leave the door open to purchase a hardware RAID controller. On my proposal to my manager, I put a footnote regarding the possible additional purchase.

I had specced out a full blown VMware, Exchange, DC, etc project, but finances aren't there to go through with it. Right now, all documents are stored on user's desktops and that is it. I feel that this is an inexpensive solution to combat the problem. The server will be backed up to an external hard drive probably nightly.

While it IS NOT the ideal solution, I feel that it is better than what I have now.. nothing. I am planning to also do one of these on a smaller scale for a branch office with 15 pc's. Our few other branch offices only have a few users at each. I figure they can back up over FTP through VPN on a weekly basis and be fine.

Dell quoted a similar spec'd server for $3400. Of course, it would have had onsite warranty, but when I do the other server for the branch location, I will purchase an additional motherboard to have on hand and a spare identical hard drive. I always keep a power supply in my supply closet. I feel that the motherboard and hard drive are the only 2 things that are likely to fail causing downtime. If a stick of memory goes bad, I'll just remove it and get it replaced.

Am I doing the right thing? If not, too late now :(. I have a fairly good feeling about it though. Once again, it's better than what we have now.
 
Just quickly glanced over everything, but you bought $90 500GB hard drives? You can get plenty of 1TB's for $90-$110
 
Yea, alot of money wasted on those HDD's. Usually its prudent to post these threads before you purchase. :)
 
Being that this is going to be an enterprise file server though, I went ahead and sprung for the enterprise class drives. Supposedly, they go through a more rigorous quality control than standard drives. I was either going to get those or the WD Black.
 
It would probably be worth it to grab a 5th drive for each and designate it as the hot spare.(I don't know the details of that raid controller). Those are the most likely to fail component.
 
bigdogchris said:
It would probably be worth it to grab a 5th drive for each and designate it as the hot spare.(I don't know the details of that raid controller). Those are the most likely to fail component.
Click to expand...

That was mentioned in my clusterbomb of a post lol. My mind cannot complete one thought without jumping to another.
 
Personally I would have gone with the dell solution, especially in dell outlet, you can snatch a cheap system and a md1000 without any real $$ spent.
 
A dell solution would have ran about $5000. $3000 for the MD1000 and another couple thousand for a server. It would have been nice, but while we grown a fine shrub, we can't get them to produce money instead of flowers.

I had my Dell premier rep quote out a Poweredge 2970, 2.5GHz quad core AMD, 16gb ram, 3 x 500gb. Cost was $3400.
 
I second buying a solution from Dell / HP / <insert your favorite hardware vendor here>. Building servers for home use is fun, building servers to run in an enterprise type environment is not. Buy something /w real 24x7 support.

Sure its an extra 2k or so, but when the mobo fails at 5pm on a Friday Dell will have one to you the next morning. How would you address this same situation in your build? How long will users not have access to all their email (your storing their .pst's) and their files while you search the web for a compatible mobo?

Do you really want to tell your boss he cant get to an important document because the server you built is down?

Just something to think about.
 
Jeff1024 said:
Do you really want to tell your boss he cant get to an important document because the server you built is down?

Just something to think about.
Click to expand...

In this case you tell the boss you worked to the budget he specified. If the money's not there it's not there. Perhaps if something like this does happen it'll make the boss think about where to spend money in the future.
 
amp88 said:
In this case you tell the boss you worked to the budget he specified. If the money's not there it's not there. Perhaps if something like this does happen it'll make the boss think about where to spend money in the future.
Click to expand...

You're boss is more sympathetic than many I have seen.
 
amp88 said:
In this case you tell the boss you worked to the budget he specified. If the money's not there it's not there. Perhaps if something like this does happen it'll make the boss think about where to spend money in the future.
Click to expand...

Yea, the boss probably wont think it like that. All he/she knows is that the file for their half million dollar project is unavailable. Take the cost and trade it for support...

Also, it probably wont be just the server you built that goes, he/she might let you take it with you though, if they are nice...
 
Jeff1024 said:
I second buying a solution from Dell / HP / <insert your favorite hardware vendor here>. Building servers for home use is fun, building servers to run in an enterprise type environment is not. Buy something /w real 24x7 support.

Sure its an extra 2k or so, but when the mobo fails at 5pm on a Friday Dell will have one to you the next morning. How would you address this same situation in your build? How long will users not have access to all their email (your storing their .pst's) and their files while you search the web for a compatible mobo?

Do you really want to tell your boss he cant get to an important document because the server you built is down?

Just something to think about.
Click to expand...

I will be building 1, possibly 2 more servers with identical motherboards. I will be ordering another motherboard (tested upon arrival) to have on hand in my supply closet. I did this with our payroll computer and it actually saved my butt one time. The motherboard went out and I had one in the closet. It has hard to pitch the selling point of redundancy/high availability, the cost to implement, and the even greater cost of NOT implementing it. I have pitched different proposals until I am blue in the face. As said before, It's not the perfect solution, but it's better than what I have no. NO redundancy for desktop storage. At least with this, I'll have Raid 10, data backed up to external media, and spare parts in the closet.

The managers have laptops, so they will just synchronize to the server instead of having all documents actually on the server.
 
amp88 said:
In this case you tell the boss you worked to the budget he specified. If the money's not there it's not there. Perhaps if something like this does happen it'll make the boss think about where to spend money in the future.
Click to expand...

THANK YOU! This is my thinking precisely. There has never been an instance in this company where data loss has occurred. In the couple instances that a desktop PC hard drive has crashed, I have been able to switch electronics on the drive and get it to spin back up. We have been very lucky. Risk vs cost of prevention. The only thing they look at is cost of prevention. Anything above $0 is too much. Sadly, the cost of risk during spring season is $450,000/day.
 
GJSNeptune said:
Like a couple hundred?
Click to expand...

It makes no sense not to spend the $30 extra now. I can spend $30 extra now to get what I will need in the future.. or spend $120 in the future to get what I could have gotten for $30 extra, now.
 
I don't think you needed either of those power supplies. Those hard drives probably peak around 10W. That power supply can power more hard drives than the Norco can hold, even with the Xeons. I'm not saying it was a bad purchase or that it's a bad power supply. The Corsair units are very good, and efficiency is important (and expensive).
 
GJSNeptune said:
I don't think you needed either of those power supplies. Those hard drives probably peak around 10W. That power supply can power more hard drives than the Norco can hold, even with the Xeons. I'm not saying it was a bad purchase or that it's a bad power supply. The Corsair units are very good, and efficiency is important (and expensive).
Click to expand...

I had actually looked for the 750TX, but it was out of stock. Hah.. I just looked and it's back in stock. If I ever go to a dual CPU set up though, I figure it'd be pushing a 650w, depending on which CPU's I choose and how many hard drive bays are populated.
 
Oh, whoops. I had it in my head that you were getting two Xeons. But even with a second one, you've got plenty of headroom. Hopefully you're not underusing the power supply. Most experts say they need to be 20-40% utilized.
 
Sasiki said:
THANK YOU! This is my thinking precisely. There has never been an instance in this company where data loss has occurred. In the couple instances that a desktop PC hard drive has crashed, I have been able to switch electronics on the drive and get it to spin back up. We have been very lucky. Risk vs cost of prevention. The only thing they look at is cost of prevention. Anything above $0 is too much. Sadly, the cost of risk during spring season is $450,000/day.
Click to expand...


Is your boss or whoever you get your budget from an IT person?
 
Sasiki said:
I will be using this server as a corporate file server for our main office, 35 users (only about 10 heavy users). All user Document folders and Outlook PST's will reside on the server. The hard drives will be configured in a RAID10 array. If VMware will work with the Intel ICH7r controller, I will do that and then Linux within the VM. I did leave the door open to purchase a hardware RAID controller. On my proposal to my manager, I put a footnote regarding the possible additional purchase.
Click to expand...

I'm fairly certain that there are zero fakeraid controllers on the VMWare HCL. It should work with them as single drives though.
 
GJSNeptune said:
Oh, whoops. I had it in my head that you were getting two Xeons. But even with a second one, you've got plenty of headroom. Hopefully you're not underusing the power supply. Most experts say they need to be 20-40% utilized.
Click to expand...

I have a 420w Thermaltake that I can stick in it until I throw more drives in it if that's the case. I'll monitor it and see how it goes.. maybe it's time to invest in a Kill-a-watt meter thing.
 
Volkum said:
I'm fairly certain that there are zero fakeraid controllers on the VMWare HCL. It should work with them as single drives though.
Click to expand...

There is some sort of file I can use during the install of VMware that supposedly allows raid to work on the ICH7r. I read that it doesn't work so well on the ich8r and 10r. Thanks for that link though. That will help a lot when / if it does come time to choose a hardware raid controller.
 
brokenarrow03 said:
Is your boss or whoever you get your budget from an IT person?
Click to expand...

Unfortunately, I am the only IT person here. I filled out a budget last year to accommodate this project as well as about 5 new laptops for our sales reps, and around 10 to 15 new desktop PC's. The sales reps are using 4.5 year old Sony Vaio's right now. Those things are nearly busted all to pieces. Some of the desktops I have around here are Celeron 900's and Athlon 2500+ (which do not perform well in Office 2007 with particular macros). Every purchase order I submit that is over a few hundred bucks gets thrown in the trash lol.

I am just trying to make do on a tight budget. May as well say, no budget at all.
 
wow dude that sucks for sure. I guess I am just baffled that they would risk $425k per day to save a few thousand bucks, but if thats where the priorities are...

I dont want to bash you by any means, but if this were my baby I would make them very clear the risk/cost. They are certainly trying to educate you how business should be run, so they should also be educated. What happens if you are on vacation? Proactive is and should be the only IT policy, not reactive.

Lastly I want to edit my comment earlier. WHEN, not if, WHEN the systems are down and the execs are losing $425k a day the last thing they will be thinking of is what they should have done earlier. It will be your ass on the line. If you say "I told you so" to them when it happens your ass will be out the door so fast...this is just a disaster waiting to happen...

Sorry if i am lecturing, I just hate to see fellow IT'ers in a mess because of poor management...
 
I would have to recommend dual PSU, UPS and an identical server that can be used for HA purposes. However, it sounds like your budget is what it is so just make sure you are watching component health closely and convince your mgmt to buy some spares (nic, hard drive, psu)
 
brokenarrow03 said:
wow dude that sucks for sure. I guess I am just baffled that they would risk $425k per day to save a few thousand bucks, but if thats where the priorities are...

I dont want to bash you by any means, but if this were my baby I would make them very clear the risk/cost. They are certainly trying to educate you how business should be run, so they should also be educated. What happens if you are on vacation? Proactive is and should be the only IT policy, not reactive.

Lastly I want to edit my comment earlier. WHEN, not if, WHEN the systems are down and the execs are losing $425k a day the last thing they will be thinking of is what they should have done earlier. It will be your ass on the line. If you say "I told you so" to them when it happens your ass will be out the door so fast...this is just a disaster waiting to happen...

Sorry if i am lecturing, I just hate to see fellow IT'ers in a mess because of poor management...
Click to expand...

All procedures will be clearly documented. My supervisor isn't a hardware guy by any means, but he can follow instructions. He is actually the programmer for our internal application. They understand the risk, but for some reason, are hesitant to do anything about it. I presented a quote for $30,000 for a HA VMware solution a couple weeks ago. They couldn't swallow it. I hate to know that I am responsible for the company's data and hate worse to know that I cannot do the necessary things to safeguard it. But like you said.. that's where the priorities are.

We are a seasonal business (wholesale plant nursery). This Spring does not look like it's going to be even with the last year, which was down from the year before. If at the end of May, we have leveled out on sales from last year, I may can sell the project. If I do, I then have this hardware to play around with. I just to get us through the next couple/few months.

I have stressed over and over to the CEO that computer equipment is a ticking time bomb. It's not IF it will fail, it's WHEN it will fail. IT is vital to the selling of our product, but is definitely no where near the top of the priority list. It really aggravates me when I see dump truck after dump truck loads of rocks being delivered for the construction of a new growing block. The cost to construct 1 growing block is more than the cost to safeguard our data, the company's lifeline.
 
Jeff1024 said:
I second buying a solution from Dell / HP / <insert your favorite hardware vendor here>. Building servers for home use is fun, building servers to run in an enterprise type environment is not. Buy something /w real 24x7 support.
Click to expand...

HP + 24X7 + iLO = happy companies and customers
 
vischo said:
HP + 24X7 + iLO = happy companies and customers
Click to expand...

I agree. I build my own servers for personal use, but the problem that comes into play with a custom built machine in the workplace is what happens when some component goes tits up that you can't get locally when needed. Downtime can cost a lot of money for each day, hour and minute, and waiting for a part to be RMA'd can take who knows how long. HP is the best OEM I've ever seen in warranty work action, a college I used to work for lost a disk on the HP 9000 on a Friday evening just before 5 pm. Because HP could not get a disk overnighted with a guarantee from a shipper that it would arrive Saturday morning, they put a technician on a plane within an hour with the hard drive in hand and he had the disk in, and the RAID array rebuilt by Saturday at 3:30 am. Thats when a OEM warranty really shines.
 
Sasiki said:
A dell solution would have ran about $5000. $3000 for the MD1000 and another couple thousand for a server. It would have been nice, but while we grown a fine shrub, we can't get them to produce money instead of flowers.

I had my Dell premier rep quote out a Poweredge 2970, 2.5GHz quad core AMD, 16gb ram, 3 x 500gb. Cost was $3400.
Click to expand...

I like my Dell reps, they work hard, but one thing they don't do is actively go out and look for the best deal for your situation (mainly cause they don't know enough.) Anyway, for about the same price you bought your system for you could have gotten a PE T605, with roughly the same performance, and more expandability and reliability minus the drives. You could always get the extra drives at newegg and it would just be an extra $200 or so.
 
Jeff1024 said:
You're boss is more sympathetic than many I have seen.
Click to expand...

I'm actually unemployed just now but it wasn't a problem at my last place. They had Dell support. In one case a drive failed, one call from the sysadmin and they sent a replacement within 3 hours. The problem is the OP can't afford this, so he's going to have to made do with what he's got.
 
Working within the limitations of your business...

you made some good choices. It all makes sense to me.

I would have purchases a Super Micro brand of motherboard, but that is just me.
 
jeff1024 said:
Do you really want to tell your boss he cant get to an important document because the server you built is down?
Click to expand...

you tell your boss to define an SLA, and come back with a quote to achieve the uptime requirement. if you're boss says five nines, you tell him to buy netapp and a generator, if your boss says it can be down for a day, five times a year, then you build your own server :p
 
Unless you have central UPS for that building you are missing one extremly important component for this system - some nice UPS - I don't know 110V ones so i can't advise which one but 1000-1500 VA unit should be below $200 and even with minimum backup time it can save company data by doing controlled shutdown on the server.before battery runs down.
 
Michaelius said:
Unless you have central UPS for that building you are missing one extremly important component for this system - some nice UPS - I don't know 110V ones so i can't advise which one but 1000-1500 VA unit should be below $200 and even with minimum backup time it can save company data by doing controlled shutdown on the server.before battery runs down.
Click to expand...

I have an APC 2200va (1850 watt) UPS on the rack. There are 2 Dell Poweredge 1950's, HP Procurve 48 port gigabit switch, Barracuda Spam Firewall 300 and Sonicwall NSA3500 router (security appliance), Nortel BCM 400 phone system, HP Proliant tower server. Utilization on the UPS is between 40 and 60% according to the meter on the front. 15 seconds after power goes off, a Kohler 65kw diesel generator fires up and the whole office is lit up.

If this additional server does cause the UPS to get near capacity, I have a 1500va I can use. I may go ahead and put it on a different UPS. Seems like all of that stuff would be drawing more than 50% utilization on the 2200va.
 
You must log in or register to reply here.
Back
Top