Jump to content
Covecube Inc.
  • 0
iceaura

Time to build a proper server

Question

Well, I've been contemplating this for years but I've only just now come across DrivePool, which is exactly what I've been hoping for! I got a little excited when Windows 8 was released and Storage Spaces looked like what I was after, however I discovered its (unacceptable) flaws after some research, then stumbled upon DrivePool!

 

First, my current system started as a large gaming/CAD workstation and has since ballooned into an always on media server too. Now, I'm really struggling for space and SATA ports (I have installed PCI to SATA adapters, PCI to USB and then to USB to SATA adapters, HDD docks, etc etc). The system runs fairly well and performs fine but I've decided to build a dedicated system with plenty of space, room to expand and lower power consumption.

 

Just for interest, the current system is:

OS: W7 Pro 64

CPU: Core i7 920 with Noctua NH U12P cooler

MOBO: ASUS P6T-Deluxe

GRAPHICS: Dual Radeon HD 5850 (running 4 screens, 3x23", 1x42")

CASE: Silverstone Raven case

RAM: 6GB triple channel RAM

An SSD for the system drive

About 10-12 mech drives ranging from 2 to 4TB - a range of brands and speeds (Seagate, Samsung, Hitachi, WD).

 

I have a 10 Gb switch running to several devices, including PC's, gaming consoles and an IP camera, wireless router (for laptops and mobile phones)

 

I've built a few PC's in the past (including this one) without any issues, but I'm no IT guy - well, nothing formal anyway, just tinkering. The idea of a server has always seemed "too professional" for me, but after some research over the past few days, I reckon I can handle it.

 

I don't like the idea of hardware RAID due to the drive limitations (I have a very multi-cultural collection of drives), which is where DrivePool will fit my needs perfectly.

 

This is my current plan for my build:

OS: W7 Pro 64 (Maybe W8 Pro 64)

MOBO: Asus H97-Plus

GRAPHICS: Use the onboard

CPU: Intel Corei5 4590

RAM: 16GB kit (2x8GB) G.Skill Ares DDR3 1600

CONTROLLER: Highpoint RocketRAID 2740 16ch

CASE: Norco RPC-4224 in a 12RU cabinet

PSU: Thermaltake Toughpower 850W

NETWORK: 24 port managed switch

UPS?

HDD: I've got a spare SSD that I'll use for the system, then transplant my mech drives over onto the controller using breakout cables

 

If I need to add future drives I will add in Highpoint Rocket EJ340 expanders - which I have a question about. Can I just keep daisy chaining EJ340's together? What happens to speed performance?

 

Also, I briefly saw something on a "landing spot" or something, where files are initially copied to an SSD before being re-distributed to the pool. This is an awesome idea and will probably end up taking advantage of this. Is there any way to kind of do the reverse? i.e. select a few files or folders in the pool and "mark" them for the landing zone because I know I will be transferring them to an external drive later or streaming them later?

 

Another question and a reason why I've been hesitant to put up a server; is there any decent "network recycling bin" type software around? I don't like the idea of deleting something over the network and then that file is gone forever.

 

So that is my build plan, what do you guys think? I've done some research, but I really don't want to have made some fundamental mistake, buy the parts and THEN realise I've messed up. Any input is appreciated, also any suggestions on a cheaper solution would be great!

Share this post


Link to post
Share on other sites

Recommended Posts

  • 0

Thanks Chris, looks like I'll be making my own plates then. How about the OS these days? Should I go to Windows 10? Windows 8? My preference is Windows 7 due to familiarity and I think I'd get the Pro version

For a client OS? 

 

Windows 10.  It's much faster than Windows 7.  And it fully supports UEFI, so boot times are much faster as well. 

Also, using Windows 10 for a few days, I'm already finding that it's MUCH better laid out, for both desktop and touch devices!

Share this post


Link to post
Share on other sites
  • 0

No, for the server. OK, I'll have to do some more research into it, I've heard it's free but there is some kind of catch? Like they might be going subscription based or something. Also, a little worried about software compatibility, although that will be less of an issue than it has been in the past (when I was into gaming emulators etc).

Share this post


Link to post
Share on other sites
  • 0

Well, as for Windows 10, it's a free upgrade if you have an activated system (for the first year, at least).  And it retains the same activation method used for Windows 10 (so if you're using KMS on your work network, then it's "stuck" on KMS). 

However, if you do a clean install (at least for a retail or OEM system), you should be able to retain the activation without any issues (just skip the key input every time it asks).

 

After next July, you'll have to buy a copy to upgrade.  

 

As for compatibility, it's mixed, but that should be no surprise.  Most everything I have works just fine on Windows 10. I've not seen anything that caused any errors (outside of add-ins for Server 2012R2 Essentials, but that was because the connector software was removed during the upgrade)

 

 

As for Server OS's, it depends on how much your willing to spend. I highly recommend Windows Server 2012R2 Essentials, as it includes all of the WHS2011 features, and then some. It includes a domain controller, and it uses "Windows Deployment Services" (a network "boot server") to load the Client Restore disk without having to have ANY CD or USB sticks connected to the system.

However, the caveat is that the OS is $400.

Share this post


Link to post
Share on other sites
  • 0

Have any of you guys that have the Norco cases had any problems with the hot swap bays? Some people were reviewing them and said they've had blow outs with the back plane. Have they resolved this issue, or would it be safer to just not hot swap and just plug in or remove with the power off?

Share this post


Link to post
Share on other sites
  • 0

Have any of you guys that have the Norco cases had any problems with the hot swap bays? Some people were reviewing them and said they've had blow outs with the back plane. Have they resolved this issue, or would it be safer to just not hot swap and just plug in or remove with the power off?

Only issues I've had with them is with a reverse breakout cable that didn't like one of the bays... It was really, really weird.

 

 

Otherwise, no, no issues at all.  And I've hot plugged more than a few drives in the system.

Share this post


Link to post
Share on other sites
  • 0

OK, its been a long time and I STILL haven't got my server up and running yet.

 

A few (minor) hardware issues - the modular cables that came with the PSU only had 2 molex connectors (and the NORCO needs 6), I also learned that different PSU manufacturers use different modular cables, which makes no sense to me. I had a bunch of spare molex cables from Thermaltake PSU's but the new PSU was a EVGA. I should have just stuck with Thermaltake. So I bought some SATA to Molex adapters. Problem solved. The other issue is that I wanted to fit 3x120mm fans instead of the 4x80mm noise machines, no problem, just drilled some holes to mount, cut out holes for airflow and mounted fans and grills both sides.

 

I've got it all put together nicely, everything boots fine, I installed Windows 8 Pro, then upgraded to Windows 10 Pro (it was cheaper this way), did all the updates, installed all drivers, installed antivirus, etc. I've thrown in 10 old spare HDDs and reformatted most of them. Then, after leaving it on for a while, it randomly reboots with the "DPC Watchdog violation" error. I found the crash would occur randomly, anywhere from 20mins to 2 hours after startup (while not doing ANYTHING on the PC). So I though that maybe it was one of the old HDDs. For the past few weeks, I have systematically unplugged and plugged in HDDs trying to find out which one may be causing the error but it seems I am chasing my tail. I'll unplug a HDD and not get a crash for over a day, then plug that HDD back in and still not get a crash. Then plug in another HDD where I wasn't getting a crash before and then getting a crash after an hour. How am I supposed to narrow down the problem? Please help, I want to get this sorted and stable before I move on to populating the server with HDDs and FINALLY buying and installing Drivepool and setting up the pools - very excited about getting to this point.

Share this post


Link to post
Share on other sites
  • 0

I'm not sure what the DCP Watchdog violation error is.  Are you sure the problem is with an HDD drive causing the crash?  It might be software.  But if it is hardware, maybe one of your backplanes is bad?  I remember you said you got the Norco 4224 (that's what I just got today).  If the backplane is the issue, Norco will replace them for you.  It might be your power supply, where if you connect one drive it's fine but 2 drives might draw to much current and something crashes.  Maybe one of your old drives is failing.. could be a number of things.  Hope you figure it out!

 

Take a look here:  http://pc.net/helpcenter/answers/windows_8_dpc_watchdog_violation

Share this post


Link to post
Share on other sites
  • 0

I am not sure if it is a HDD. Or if it is a hardware or software error. That's the problem - I can't narrow it down to isolate the issue. I have a feeling it might be an OS/driver issue but I don't want to roll back to Windows 8 (yuck). If one of my old drives were failing, I can't find which one because one day it is fine, the next it seems to be causing a crash and then another day it might seem to be another drive. I haven't managed to narrow this down to any particular drive.

 

I wouldn't think it would be a power draw issue from the drives - they wouldn't even draw 10W each x 10 drives = 100W. The PSU is 1000W, unless there is a limit on each cable? I have one cable going to 2 of the rows (8 drives) and another going to 4 rows (16 drives). I only have 10 drives connected mostly to the top two rows.

 

I am going to check the device manager when I get home (I keep forgetting to do this) and also check that my SSD firmware is up to date. Although, why would it crash after a random period?

 

Also, just for fun, my other 2 PCs are having issues at the moment:

PC1 - Sometimes youtube videos will go stuttery and the audio will distort (both IE and Firefox). Whenever I start up or wake it up from sleep, it will hang and I have to hard reboot, then skip the startup repair and it will boot.

PC2 - One of the graphics cards fans has been making terrible noises the last few months every now and then (I have only just narrowed it down to the fan, have ordered a replacement).

 

I've been avoiding these problems but I really need to sort them, most importantly, the server.

Share this post


Link to post
Share on other sites
  • 0

I never upgraded to Windows 8, and am definitely staying away from Windows 10.  It's a personal thing with me.. I gotta have my Media Center.  :)  Anyway, since I've not upgraded, I don't know what utilities Windows 10 (or 8) has available to test the hard drives.  They've probably got something, like chkdsk.  I would run that.  Maybe you have a bad sector (or more) and when the computer is trying to read or write to it, that's when it fails.  I had that problem once, and couldn't identify it.  But after figuring out that the drive was failing with bad sectors, it went away after I replaced it.

Share this post


Link to post
Share on other sites
  • 0

Yeah, I would've preferred to stick with Windows 7, but W10 looked good and it was MUCH cheaper, so I went with it. I can only go back to W8, which I haven't used much of but what I have seen is terrible, so I am kind of stuck with W10. I might try running chksk on all the drives but that will take a very long time, I already ran a check on one of them with no errors, it took hours and hours (and that was a small drive), so I thought it would be faster to just check the drives by plugging/unplugging them to reproduce the error - which I haven't been able to do repeatably. It shouldn't be trying to read or write any of the drives though, as I start the computer up and don't touch a thing, then it will crash some random time later.

Share this post


Link to post
Share on other sites
  • 0

OK, its been a long time and I STILL haven't got my server up and running yet.

 

A few (minor) hardware issues - the modular cables that came with the PSU only had 2 molex connectors (and the NORCO needs 6), I also learned that different PSU manufacturers use different modular cables, which makes no sense to me. I had a bunch of spare molex cables from Thermaltake PSU's but the new PSU was a EVGA. I should have just stuck with Thermaltake. So I bought some SATA to Molex adapters. Problem solved. The other issue is that I wanted to fit 3x120mm fans instead of the 4x80mm noise machines, no problem, just drilled some holes to mount, cut out holes for airflow and mounted fans and grills both sides.

 

I've got it all put together nicely, everything boots fine, I installed Windows 8 Pro, then upgraded to Windows 10 Pro (it was cheaper this way), did all the updates, installed all drivers, installed antivirus, etc. I've thrown in 10 old spare HDDs and reformatted most of them. Then, after leaving it on for a while, it randomly reboots with the "DPC Watchdog violation" error. I found the crash would occur randomly, anywhere from 20mins to 2 hours after startup (while not doing ANYTHING on the PC). So I though that maybe it was one of the old HDDs. For the past few weeks, I have systematically unplugged and plugged in HDDs trying to find out which one may be causing the error but it seems I am chasing my tail. I'll unplug a HDD and not get a crash for over a day, then plug that HDD back in and still not get a crash. Then plug in another HDD where I wasn't getting a crash before and then getting a crash after an hour. How am I supposed to narrow down the problem? Please help, I want to get this sorted and stable before I move on to populating the server with HDDs and FINALLY buying and installing Drivepool and setting up the pools - very excited about getting to this point.

 

 

Sorry if I missed anything else here.....

 

 

On the power supply you're using... check the "DC Output" stats for it. Specifically, look for the +12V section. This needs to be at least 2A per drive you're connecting.  Since it sounds like you have a Norco RPC-4224 or 4024, that means that this should be 50A or higher (48, plus internal drives or other components). This may be broken up, depending on your power supply, and if it is, then it should be "split" between the drives accordingly.

 

(Why 2A per? Because that's the startup power pull for most HDDs, and the highest value rated for most HDDs)

 

If it's not up to this rating, the power supply may be failing to provide enough power to the system and that could be causing the BSODs.

 

 

 

I would also recommend running a memory test on the system, as that may indicate the issue.

 

Otherwise, upload the dumps to us, and we'll take a look.

http://wiki.covecube.com/StableBit_DrivePool_System_Crashes

 

But it could be something such as a bad driver on the system.

Share this post


Link to post
Share on other sites
  • 0

I'm positive it's not a power issue, I only have 4 drives connected and it is still crashing, this latest one it crashed after being on for over a day without me doing anything on it. I had a look at device manager and there is an exclamation mark next to the audio devices and I do remember looking at that after I first built it and turned it on, I tried looking for replacement drivers online but gave up as I had other stuff to do and then I forgot about it. I do still get audio though.

 

I'll have a look at the dumps and run a memory test. On that note, what memory and hard drive tests do you recommend, I know I have a few programs but I would like to know what you guys are using

Share this post


Link to post
Share on other sites
  • 0

Do you need audio on a server?  See if you can disable it in your BIOS.  Without drivers, that "might" be something?  I know on my server, I have no use for Audio, so I have it disabled in my BIOS.  Windows doesn't even detect any audio hardware then. 

 

As for memory and hard drive tests, I use Memtest86 but Microsoft has one built into Windows.  I think you access it through boot menu (F8 before Windows loads), though I'm not sure it's included or not in Win8/10.  As for hard drives, I have LOTS of programs.  I highly recommend getting Hiren's Boot CD (link).  Every tinker-er should have it.  Burn the CD then enable your BIOS to boot from it.  It has Windows, DOS and Linux utilities to access various parts of the computer at low-level, and they're all completely free.  They're just all collected on 1 disk with a realy easy menu to navigate to what you want.  They have so much good stuff in there, like AntiVirus, HDD test and repair tools, Drivers, Backup and Recovery, and lots more.  You can also run it in Windows and see if there are helpful programs you can run from there too.  Also, I just checked, and they have Memtest86 on there as well!  The HDD tool I'd recommend is HDAT.

Share this post


Link to post
Share on other sites
  • 0

I'm positive it's not a power issue, I only have 4 drives connected and it is still crashing, this latest one it crashed after being on for over a day without me doing anything on it. I had a look at device manager and there is an exclamation mark next to the audio devices and I do remember looking at that after I first built it and turned it on, I tried looking for replacement drivers online but gave up as I had other stuff to do and then I forgot about it. I do still get audio though.

Ah, okay, then that may not be a power related issue then.  

If you can run just the system disk and no other drives, does that work?  (the reason I ask, is that I've had a single shorted out drive that would cause the system to hang and BSOD, so just in case). 

 

But as hansolo77 mentioned, disabling the audio and other devices in the BIOS may be a good idea.

 

 

 

I'll have a look at the dumps and run a memory test. On that note, what memory and hard drive tests do you recommend, I know I have a few programs but I would like to know what you guys are using

 

As for testing tools, the built in memory diagnostics tool work great. Though, the extended test does work better.  

You can also access it inside of windows (but it wants to reboot to run it).

 

As for disk tools .... shameless: StableBit Scanner.  If you right click on a disk, there should be a "burst test" option. Running that is a good way to stress test the drive/cable/controller, and see if that's the issue.

 

 

As for boot disks, I prefer UBCD. Much less but still a lot.

Share this post


Link to post
Share on other sites
  • 0

Thanks guys, I figured memtest86 and I have used UBCD too. I ran a surface scan using Easeus Partition Master I think. Haven't used Hirens Boot CD, might have to check it out. I would use StableBit Scanner but I want to get the server up and running before I start a free trial to make sure everything works out - I'm pretty sure it will  ;) . I have tried only running the system disk (the other two SSDs are connected but not initialised) and it didn't crash after a day but as I recently found out it CAN crash, even after a day of stability, so I will probably try again with just the system disk. The only problem is, I don't know how many days before I can safely say that it is stable and that the problem has to lie in the mech drives/controller card. I'll probably give it a week.

 

I probably won't need audio but I'd rather have it just in case and would rather have it and not need it than the other way around, so I should try and get rid of that exclamation mark. If I still get crashes running only the system disk, fixing the driver/disabling the device will definitely be my next focus.

Share this post


Link to post
Share on other sites
  • 0

Something else to consider... is your server going into standby at all, or is it on 24/7?  It might be crashing when the system goes into standby, or wakes up.  Just a thought I just had, might not apply.

Share this post


Link to post
Share on other sites
  • 0

I turned off the power saving settings, so it won't be that, it's set to run 24/7 (even the monitor for now). It just randomly crashes when it feels like it (10 mins, 20 mins, 2hrs, 1 day, etc) without any apparent reason - this makes it real fun trying to figure out the issue, lol. I'm not sure about this, but is there any way that being connected to the network/internet may have influence on it crashing?

 

On another note, I upgraded one of my other PC's to W10 and had no audio at all (so I reverted back). Seems W10 has real audio driver issues <_<

Share this post


Link to post
Share on other sites
  • 0

I have had a look at the event logs (I'm uploading it now but it's 600MB (!)), I'm not too sure what to look for though. I see warnings related to the disks (I think all of them) such as "Disk 9 has the same disk identifiers as one or more disks connected to the system" and "An error was detected on the device\Device\Harddisk during a paging operation". So it seems that it has something to do with the disks but I believe it will be the controller card causing it.

Share this post


Link to post
Share on other sites
  • 0

Have you tried disabling the automatic rebooting option? 

  1. Right-click on My Computer and choose Properties.
  2. Click on Advanced System Settings on the left hand menu.
  3. Click on the Advanced tab and then click on Settings under Startup and Recovery.
  4. Uncheck the Automatically restart box under the System failure heading.
  5. Click OK and close all the previous windows.
 

Also, do you have a "C:\Windows\MEMORY.DMP" file, or any contents in the "C:\Windows\Minidump" folder? If so, upload that to us:

http://wiki.covecube.com/StableBit_DrivePool_System_Crashes

Share this post


Link to post
Share on other sites
  • 0

I'll have a look at that setting when I get home. As for the memory.dmp, I tried to upload it but I guess it was too big (600MB). Can I start a new memory.dmp file somehow? (I haven't looked into doing this yet)

Share this post


Link to post
Share on other sites
  • 0

If you're getting a BSOD and your system reboots, try running NirSoft's BlueScreenView (get it here).  On my system, it was kinda hit or miss.  It was pointing to IRQ issues, Memory Issues, Drivers, etc.  In the end, I think my problem is my bad PSU.  All the BSODs I was getting were just from whatever the system was doing at the time.

Share this post


Link to post
Share on other sites
  • 0

Upload memory dumps here:

http://wiki.covecube.com/StableBit_DrivePool_System_Crashes

 

Also, you may want to zip up the file (right click, send to -> Compressed folder) first. This should greatly reduce the size.

 

 

And as for new memory dump files... the "C:\Windows\MEMORY.DMP" file is overwritten each time the system BSODs.  The minidumps should be a more "historical" collection.

 

ALso, you can run "BlueScreenView" or "WhoCrashed" to view these yourself, if so inclined.

Share this post


Link to post
Share on other sites
  • 0

OK, I'm pretty sure it is something to do with the controller card now. I looked at the event viewer after the crash, it was saying those same error messages mentioned earlier for HDDs 2, 4, 6, 10, 13, 15. I looked these up in the disk management. They are strange little 16MB drives that I'm guessing are created due to the controller card. What do they mean?

 

I checked the memory.dmp file directly after a crash, it was 585MB, I've compressed it to 110MB and am trying to upload it now. I should have the time for it to complete the upload today.

 

By the way, thank you for all your help so far, I really do appreciate it :) Sorry I'm a bit sporadic in my troubleshooting, I rarely get the time to sit down and check on it but I am still really keen to get this solved.

Share this post


Link to post
Share on other sites
  • 0

Yeah, it's definitely the controller.

 

Looks like the 3ware driver is causing the crash specifically. 

 

 

Are you using the "Unsafe DirectIO" option in StableBit Scanner? If so, this could be causing the crash, as it may be be sending commands to the card that could cause this (that's part of why it's "unsafe").  

If you are not using the "Unsafe" option, then it sounds like there is an issue with the card. Either updated the drivers, or ... replace the card.

 

 

 

As for the drives, I'm not sure why you'd see those, but I'm not that familiar with the controller card in question.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Answer this question...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

×
×
  • Create New...