Server keeps shutting down (sbs 2008)

Well I thought the issue was solved; the server stayed on all weekend, then out of the blue this morning, it is down.
Ok, with that being said, I was thinking of purchasing SISoft's Sandra to dp some hardware checking, but not sure what types of test to run, and what to look for in the results. Can someone who may or may not have used this software tell me how good it is, recommend one that is good. Also please hlep me with what type of tests to run and what to look for in the results. I hope or assume that the software will show some kind of red flag is something is not right.
xzay1967Asked:
Who is Participating?
 
bclongacreCommented:
I would double check to make sure that your processor heatsink fans are running properly/smoothly.  I say this because we just lost a sleeve bearing in a fan, and the only thing that triggered us to know specifically what it was, was a slight bumping/knocking sound.

If you are running Dell servers, or another manufacturer similar to them, they may have Open Manage, or a similar piece of software running on them, that can report the status of components like that, but as our server was a custom build, we did not have that option.
0
 
ZombieAutopsyCommented:
Have you looked at the event logs in windows? That can usually give you a pretty good idea of why the system is shutting down.
0
 
Jonathan_OXLCommented:
You can find the reason why your server shutdown unexpectedly in the event log like ZombieAutospy said. If it doesn't give you anything then try to monitor your system resources like CPU and harddisks.

Good luck!
0
Cloud Class® Course: Python 3 Fundamentals

This course will teach participants about installing and configuring Python, syntax, importing, statements, types, strings, booleans, files, lists, tuples, comprehensions, functions, and classes.

 
xzay1967Author Commented:
The event log only system was shutdown unexpectedly at <time>. It is an HP Proliant ML310.
Now the darn thing won't even turn on. I am begining to think it might be the power supplies; there are two of them on there. Even if one is bad, shouldn't the system still come on? Kind of hard to think that both of them are bad at the same time. As for monitoring the hardware, I want to do a scan, but not sure what to look for with the sandra software.
0
 
bclongacreCommented:
That depends on the configuration.  Some servers use 2 x power supplies for required power, and full redundancy would be 2 x 2 power supplies, while others only require 1 so in that case redundancy would be 1 x 1 power supplies.
0
 
bclongacreCommented:
Have you tried changing the UPS the server is plugged into?  I also encountered an instance once where, a battery backup, was pushing adequate voltage, but if it dropped to battery the output sagged and the servers dropped off.
0
 
xzay1967Author Commented:
Ok I will try switching to direct power for a bit and see if it comes on.
0
 
xzay1967Author Commented:
Once I plugged the sever directly into the wall, it came right on up. I purchased a new UPS, now I will monitor and see what happens the next few days.
0
 
bclongacreCommented:
glad to hear your back up for now, hope she holds together for you.
0
 
xzay1967Author Commented:
Well guys, it stayed up about 24hrs, then shutdown again......arrrrg. Now I have it on direct current to see how long that would last. Although I think it might do the same thing being that it was plugged into a new UPS.
0
 
Cliff GaliherCommented:
sounds like you are purchasing underpowered UPS's. A UPS can only deliver so much power, and consumer UPSes can't keep up with the power drain of multi-core CPUs, multiple high-speed disks spinning in a RAID array, and multiple power supplies trying to pull enough power to monitor themselves and each other. Make sure your UPS is rated to provide as much power as your server is pulling. A small power meter is essential for this type of work. I have a Kill-a-Watt on my bench for exactly this type of stuff.

Then, once you know what you need and you've adjusted for some overhead and growth, buy a server-class UPS. If you can find it in Best Buy, it isn't ready for your server.

-Cliff
0
 
xzay1967Author Commented:
I hear what you are saying, but why would the original UPS work for almost two years, then all of sudden not provide enough power (if the UPS is the issue)? The new one I bought is a 2000VA from compusa. I will leave it on direct power and see how long or if it stays on.
0
 
Cliff GaliherCommented:
Server load increases, hence my comment to "plan for growth."  The simple truth is that all of this is speculation. I am basing this on the fact that you could not turn on the server when plugged into the old UPS but moving to direct power seemed to *IMMEDIATELY* resolve the issue. That screams UPS...and I'm not one to believe in coincidences.

Then you move to new UPS and problem reappears.  Tells me that your server is pulling more power than your UPS choices have been able to provide. So there is my logic behind my recommendation, but as I said, cannot be definitive without extensive troubleshooting and burn-in.

-Cliff
0
 
xzay1967Author Commented:
You make a good point Cliff. You mentioned that you had a kill-a-watt, but googling gives me stuff about killing etc. I assume that the kill-a-watt minitors your server's power out put and gives you some kind of alarm. Thiis is the new UPS I bought. I will try to find some kind ogf tool that can tell what my server is putting and match it against the ups.
0
 
xzay1967Author Commented:
I put my server specs into the APC wizard and it generated the recommended UPS's in the screenshot. So I compared their recommendations to the new one I boght and I exceed the specs. Oh one other point if I may, even if the UPS I had was under power, as long as it is not on battery backup mode, would that still be an issue? Please I assure you, I am not trying to argue to disagree with your input or advice, I am just trying to get a better feel for the situation. Oh another point (if it matters), this is a refurbished server from HP, so I am trying to run aground the reason for the initial cause for it to be returned and then refurbished. Maybe the prior owner had the same issue, and now it has come back to haunt me, lol.
0
 
xzay1967Author Commented:
Forgot screen shot. APC recommendations
0
 
Cliff GaliherCommented:
Not all UPS's are created equal. Your off-the-shelf UPS is simply not designed to handle the power draw that a server requires. There is a reason why even APC has their "Back UPS" line for workstations and their "Smart UPS" line for servers. The Back-UPS line goes up to 1500VA, but doesn't have the "clean" power and other specs that servers typically need to be stable.

Example of an APC UPS that I'd never use on a server, but technically exceeds the specs of your screenshot:

http://www.apc.com/products/resource/include/techspec_index.cfm?base_sku=BR1500G&total_watts=200

As far as being off battery power, to properly account for brown-outs and such, they are always monitoring the power draw (higher end UPSs are actually always on, regulating power and you aren't on the grid directly at all). If you are pulling more power than spec'd, or if you are pulling power and the power is a little dirty, it causes the UPS to behave strangely and can actually cause problems.
0
 
xzay1967Author Commented:
aThanks cgaliher, I understand what you are saying, but as I mentioned in my prior posts, they initial UPS worked fine on the server for over a year. When the problem began a week or so ago, I went and bought another UPS (http://mail.ultraproducts.com/product_details.php?cPath=35&pPath=154&productblD=448) and this one out specs the initial one. The problem still occurred. In the screenshot from APC, I ended up with that after plugging in my server info into their wizard. My server is a HP Proliant ML310 and two power supplies, and a mirror raid. Well so far the server is up and running on direct power. I am giving as few more days on direct power to see if it stays on before ruling out everything else. What device can I use to measure the power draw to determine the correct rating UPS to buy?
0
 
bclongacreCommented:
here are a couple of links you can use to test what the real voltage/amperage draw is.

http://www.ehow.com/how_4927249_measure-transformer-current-draw.html

http://www.ehow.com/how_5328646_calculate-amperage-draw.html

Just curious, have you checked inside your power supplies, do you have any "popped" or bulging capacitors?  If so, everything may appear fine, but the capacitors may no longer function as required for stable operation, but not poor enough for it to be declared a failure, this could possibly be exacerbated as they become warmer.  If you have bad capacitors, if you are comfortable replacing them then it is a very cheap soloution, so long as you can sodar, otherwise, taking them (the power supplies) to an electronics shop, IF they appear to be bulging at the top, which should be flat.  The other option is to purchase a replacement power supply, so you can swap them one at a time to be repaired, if needed.  Having spare parts never hurts, except for the initial purchase outlay.
0
 
Cliff GaliherCommented:
The APC wizard should be good enough, but if you want, pick up a power meter such as a Kill-a-Watt. As far as specs go, like I mentioned, specs that you see on a website aren't everything. There is a reason that the wizard recommended the Smart-UPS line instead of the Back-UPS line even though both have similar "specs." Some things just don't show up well on paper, and the Smart-UPS line is the proper choice (as the wizard indicated) for a server.

Good luck,

-Cliff
0
 
xzay1967Author Commented:
Currently the server is on direct current, it has been like this about a week now. Since being put on DC, it has not shutdown once. I am going to leave it on there a few more days, then buy a smart ups and see if it keeps the server one.
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.