Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations Mike Lewis on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Need some EXPERIENCED feedback 2

Status
Not open for further replies.

OneMadCoder

IS-IT--Management
Mar 26, 2008
59
I don't know if this the right forum, but I couldn't find anywhere else to post it.

I know everyone here has enough experience and the background to answer my question. I recently had a situation where my expertise was in question.

Our company has a (so-called) server running Windows Server 2003 Enterprise. It's main services are MsSQL, MSSQL, Flash Media Server, Exchange are just a few. It currently has a user base of 84 users, mostly only a portion of them are active. Although 80% of them use exchange. All but 8 of them are remote users, some of which use VPN's.

Recently, the server has been bombing. BSOD's and a Tape Drive hardware failure which I replace with a 1 Terabyte external backup drive and a redunant off-site (Cozy.com) System Partition backup.

Sicne I am fairly new to the company, It seems the owner is blaming me for the recent problems. I have close to 15 years experience in Network Administation (6 of those for AT&T)and a handful of certifications. I explained that this machine is not powerful enough to handle the current configuration and it's user base and a new real server
is needed. Not believing me he wants to fly in some 21 year old kid with NO network experience who develops the company website to see if a new server is actually needed. This has made me beyound furious.

So now I'm asking you folks. Is this machine enough to handle the services described above.

Server specifications:
HP ML310

Processor type:
Intel Pentium 4 Processor (1.86 GHz, 1066 MHz, 2MB L2)

Processor cores:
Single

Processor cache
1 MB Level 2 cache

Multi-processor
1 (one CPU supported)

Max front side bus
1066 MHz

Memory type
2 Gigabytes (PC2-2100 DDR2)
400mhz

Max memory
4 GB

Hard Drive
(2) 160gb Sata Drives

Media Drive:
16xCD-rom

Raid Controller:
Integrated SATA RAID 0, 1

Operating System
Windows 2003 Server Enterprise
 
After re-reading my post I noticed a ton of typo's.Please overlook them. I am educated, just not very sane at the moment.
 
I'd be surprised if 84 people were able to crash the machine just through 'overuse' but then that depends on so many things. How much dbase work are they doing? How much email is going through the machine.
Enterprise machines are for the enterprise-I know places where thousands of users connect to a single machine-probably with only slightly higher specs than yours. Still it depends on how much 'stuff' your users are doing.
Have you checked perfmon to see if the machine is handling things well?

BSOD's could be to do with applications rather than the OS or the machine itself.

How old is the machine? Are the drives ok? What does eventviewer say when the crashes happen?

Here's a few links as to whether your machine can handle things:


And using perfmon

Performance Advisor may help too:


Hope that helps.
 
Thank you, you made some very great points that I will check in to tomorrow.

I've got all of Sysinternals apps (now MS) but honestly it has slipped my mind to use them to diagnose any problems.

Again, Thank you.
 
If you can, give us some examples of the BSOD screens and some of the event ids and we can then see if we can help you out.

Good luck.
 
Are the BSODs that you are receiving consistent in nature? Can you perhaps post some of them or look them up? I agree with Elmurado that server's bombing are not caused by user overuse but more like abuse. I would check the event logs around the time of the BSODs occur and then start from there.
HTH
 
Machine seems fine to me for its purpose...are there bigger and better machines available? Sure...but don't get blinded by that. Use your experience, diagnose the problem, and fix them.

It would be nice to start over from scratch with a machine you built...but it would be better if you found the source of the problem and fixed them.

I'm Certifiable, not cert-ified.
It just means my answers are from experience, not a book.
 
The BSODs are not caused by the number of remote users involved ( likely bad/conflicting driver) but with 84 users, all the resident programs, and the low grade raid I would upgrade the machine.
If upgraded I would not place Exchange on the new server. As to hardware I would get at least a duo core close to 3.0 Ghz, with 4 gig min ram, I would go 8 Gigs ram with a 64 bit OS, along with an SAS raid sub system, OS on a raid 1, data on a separate raid 1 or raid 5.
With 84 remoting users, the machine should basically be a TS with very few resident programs. The original creator of this server expected way to much from this machine, Terminal servers demand far more resources than a normal server.
If the Boss object to the upgrade expense, with the number of remote uses, the speed increase will pay for the server in employee efficiency.


........................................
Chernobyl disaster..a must see pictorial
 
Its probably choking to death on its lack of RAM or at least causing excessive swaping. MSSQL and Exchange love RAM and 2 gig just isnt enough, i would get that up to 4 gig right away. Could be a heat issue also, its obviously an older box and the cpu heatsink is probably full of dust bunnies or you may have a fan out that would lead to heat build up and a crash.

Just a couple of thoughts,



RoadKi11

"This apparent fear reaction is typical, rather than try to solve technical problems technically, policy solutions are often chosen." - Fred Cohen
 
Oops, yes, misread the spec, definately fill it up with RAM first, clean it out while you're in there, then start the diagnosis.

I'm Certifiable, not cert-ified.
It just means my answers are from experience, not a book.
 
I would suggest that you monitor the server, who's to say the amount of RAM, the CPU type speed, etc. is at fault, as preformance needs to measured based on your site (One's site is different then anothers). You can go to microsoft's web site and check what the minimal requirements are to run the application. Don't get me wrong those specs seem to be low for all that you described however without dusting off the preformance monitor and actually checking for high-utilization of several differnet physical properties, then you can't just say it needs to be upgraded... You need to prove that there are preformance issues. If you aren't seeing high utilization in processor %, virtual memory utilization, disk queue length etc. then maybe your problem is something else, i.e. application issue, exchange issue, use the event log to begin your search if you find preformance is not the issue.

regards,
Dave
 
BSOD is almost always a hardware or driver issue. Not enough computing power would manifest it self in slow performance.

I would back track the changes and upgrades made that relate to when the problems started. Also do a mem test and make sure those modules are still good. I have learned things work until they dont.

The specs do sound a little low to run all that stuff, but that is a separate issue, it should run. Something is causing the OS to crash.
 
With RAM so cheap, I'd replace all of it with new stuff, packing it all the way up to the 4GB.

I'm Certifiable, not cert-ified.
It just means my answers are from experience, not a book.
 
Star to mantastic36 for providing what I consider to be the best comment to date on this.

Is the server hardware capable of handling 84 users? PROBABLY... but you have so much going on on that server, who can be really sure? If it were JUST an exchange server, then DEFINITELY. If it were JUST a file server, then DEFINITELY. If it were JUST a domain controller, DEFINITELY... but SQL is something - who knows. Is the SQL database in charge of a web site with constant updates? And which one - the MySQL or the MSSQL or both?

The machine is PROBABLY just fine, spec wise... but I'd still replace it now. Consider this - the hardware in the system sounds old - 3+ years at least. And in my opinion, since this server CLEARLY runs your business, you should have a server that is under an appropriate warranty.

Now, assuming your boss doesn't value the ability to sleep at night, knowing the server is protected under warranty, I would agree with others posting - BSODs are almost certainly not caused by the users and they are almost certainly not caused by you. You need to post the KIND of error messages - STOP 0x0000007B for example? You should look for minidump files and analyze them. The BSODs are probably because of a bad driver update OR because of a failing piece of hardware.

And yes, I'd probably suggest upgrading the RAM... but is it, strictly speaking, necessary? Who knows. You could find out with some proper analysis, but it wouldn't affect the BSOD problem unless you by chance have bad RAM and the upgrade inadvertently replaced the bad RAM. You can control RAM usage in MySQL and MSSQL (though not in Exchange).

-Lee

Those who ask why, learn
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top