Building this server for someone else, and I would like it to be completely reliable before I ship it off.  First, the specs:

 

(basically a 15 drive budget box with a healthier budget)

Mobo: Biostar A760G M2+ (known good)

CPU: Sempron 140 (known good)

RAM: 2 GB Kingston ValueRAM DDR2 800 (known good)

PSU: Corsair CX500 (new)

SATA Cards:

  Supermicro AOC-SASLP-MV8 (new)

  PCIe x1 SIL3132 (known good)

Cables: SAS-SATA Forward Breakout x 2 (new), generic SATA (new)

Power Splitters and adapters: Molex to 2xMolex x 3, SATA to Molex x 3 (new)

Drive Bays: Icy Dock MB-455SPF 5-in-3 x 3 (new)

Drives: precleared 2 TB WD EARS w/ jumpers x 8 (new)

USB Drive: Lexar Firefly preconfigured with unRAID Pro license by LimeTech (new)

 

Currently running unRAID version 4.6-rc5.  I've also tried 4.6-rc3 and 4.5.6 with the same results.  No add-ons installed except for the powerdown script.

 

Photo here.  I have to say that this is the prettiest server I've built so far.

 

First off, I've configured this server to use a static IP, as I've found that static IPs are more reliable on my network.  I have three static IP unRAID servers on my network:

192.168.0.100 (my personal server)

192.168.0.101 (my dev server)

192.168.0.111 (this 15 bay server)

 

All set to 'Workgroup'.  Only my personal server is the local master.  All the other network settings are the same across all three servers:

Obtain IP automatically: No

Netmask: 255.255.255.0

Gateway: 192.168.0.1

Obtain DNS automatically: No

DNS Server 1: 192.168.0.1

DNS Server 2: <blank>

DNS Server 3: <blank>

 

As a bit of backstory, when I first built this server I was using a new PCIe x1 SIL3132 card that turned out to be defective.  It caused crashes on boot and all kinds of weirdness.  I then replaced it with a known good one from my dev server, and I haven't seen any of those same problems.  I'm sure the original card is defective as it makes my dev server crash as well.

 

The current issue is centered around the unRAID web interface.  Sometimes it is snappy and responsive as it should be, and sometimes it hangs, takes forever, or simply stops working.  I haven't noticed any pattern that causes this; it is frequent, yet intermittent.  It could be something simple like refreshing the page, or it could be when assigning a new drive, or starting or stopping the array.  When it stops working altogether, sometimes it will fix itself if I wait long enough, other times I'll have to reboot the server (from the console) to get it responsive again.  Browser is Chrome on Win7, haven't tried a different one yet.

 

From the system console the server seems to behave normally.  However, all I've done from the console is preclear drives, reboot/powerdown, and run initconfig.

 

Last night I installed all the drives, assigned them (which was painful as the web interface kept crashing), and started a parity sync.  Attached is the syslog from this morning just after the parity sync finished (time was normal - less than 10 hours).  I'm now kicking off a parity check and heading to work.

 

I have a feeling that this is an issue with my network and not with the server, but I would like to be absolutely sure before I ship this thing off.  I don't see how bad hardware could cause this kind of behavior, unless maybe it is a bad NIC or something like that.  I've swapped out my network cables to no avail.  I don't have another NIC I can try at the moment, but I could try another motherboard if necessary.  My router is a D-Link DIR-655.

 

Also take a look at these screenshots of the common crashes I keep seeing.  In the second one, notice the images on the page not rendering, and the 'website cannot be found' error at the bottom.  Whenever this happens, the button I just pressed (refresh, start, stop, check, etc.) won't work and I'll have to refresh manually and press it again.

 

Screenshot 1 (when trying to access the syslog)

Screenshot 2 (when pressing refresh)

15baysyslog.txt

Do you always use 0.111 for the servers you build?

 

If so, then you may be running into an issue where your router is doing some stuff for you.  My Cicso/Linksys running DDWRT tries to help me by associating a host name with the IP address.  If you set up another server with the same IP and a different host name your router might be trying to help you.

 

Try cleaning the router (I can revoke an IP which will force the computer to re-associate) and it should show up correctly

No, actually, this is the first time I've used 111.  Generally I use 101, but this time around I wanted to be able to use my dev server to preclear drives while working on this new one.

 

However, when I first built this server the IP was 101, so maybe the router is still confusing it somehow (one host name with two different IPs, or something like that?)

 

I will try cleaning the router, thanks for the suggestion.

So this is bizarre - the server in question doesn't even show up on my router's lists.  I tried changing the static IP to .0.105 and it still doesn't show up.  I don't know how to revoke the IP if the IP doesn't show up in the list.  Any more ideas?

So this is bizarre - the server in question doesn't even show up on my router's lists.  I tried changing the static IP to .0.105 and it still doesn't show up.  I don't know how to revoke the IP if the IP doesn't show up in the list.  Any more ideas?

Is it possible you have the same MAC address in multiple network.cfg files on multiple servers. They'd all be trying to respond and likely none would.

So this is bizarre - the server in question doesn't even show up on my router's lists.  I tried changing the static IP to .0.105 and it still doesn't show up.  I don't know how to revoke the IP if the IP doesn't show up in the list.  Any more ideas?

Is it possible you have the same MAC address in multiple network.cfg files on multiple servers. They'd all be trying to respond and likely none would.

 

Joe, you are absolutely right.  How do I fix this?  Or, is it even a problem since this new server won't be on my network once I ship it off?

So this is bizarre - the server in question doesn't even show up on my router's lists.  I tried changing the static IP to .0.105 and it still doesn't show up.  I don't know how to revoke the IP if the IP doesn't show up in the list.  Any more ideas?

Is it possible you have the same MAC address in multiple network.cfg files on multiple servers. They'd all be trying to respond and likely none would.

 

Joe, you are absolutely right.  How do I fix this?  Or, is it even a problem since this new server won't be on my network once I ship it off?

You can delete the HWADDR line from the config/network.cfg file on all of the servers (and then reboot them)

I deleted the HWADDR line from both servers and rebooted them (I used Notepad, was that a mistake? I figured it would be OK since I wasn't adding any lines).  After the reboot I confirmed that the line is still gone from the network.cfg files.  However, when I view the settings page on each server, the MAC address is still the same!  Shouldn't it have generated a new one?  These servers are using completely different hardware (different motherboards, NICs, etc).  I don't understand how they could have the same MAC address, since I thought the MAC address was generated based on the hardware so that it would always be unique.

 

The router continues to show just one of the servers as well.

 

However, I will say that since doing this I haven't seen any crashes or hangs while clicking around in the web interface.  So the symptoms appear to be gone.  Is it safe to call this one solved, even though weirdness still persists?

I deleted the HWADDR line from both servers and rebooted them (I used Notepad, was that a mistake? I figured it would be OK since I wasn't adding any lines).  After the reboot I confirmed that the line is still gone from the network.cfg files.  However, when I view the settings page on each server, the MAC address is still the same!  Shouldn't it have generated a new one?  These servers are using completely different hardware (different motherboards, NICs, etc).  I don't understand how they could have the same MAC address, since I thought the MAC address was generated based on the hardware so that it would always be unique.

 

The router continues to show just one of the servers as well.

 

However, I will say that since doing this I haven't seen any crashes or hangs while clicking around in the web interface.  So the symptoms appear to be gone.  Is it safe to call this one solved, even though weirdness still persists?

The router still knew them by the hardware addresses.  It had to be rebooted too as well as anything on the LAN that was caching the translation.

Thanks Joe.  I've rebooted everything - the router, both servers, and the desktop PC.  MAC addresses are still the same, and the router is still only showing one of the servers (although this time it is the new 0.105 one as I booted it up before the 0.100 one).

 

However, since there don't seem to be any of the problems I described earlier in this thread, can I just let this one rest, or is there something that will come back to bite me (or the person who is getting this server)?

You should be fine and this should not happen to the customer getting the server.

 

If you want to do one more test:

remove the HWADDR line from both files (should still be since you already did that)

run the files through fromdos just to make sure nothing windows related got put in (use NotePad++ to edit files from now on)

shut the servers down

in your router settings remove any remanent of the HWADDR that was listed.

 

Hopefully that will fix your issues.

if i'm not mistaken MAC address are not calculated but assigned to each individual network card -

 

also your router would only see your server when it tries to access the internet or other computers on your network

 

since your router didn't give the address to the unraid server it only recognize it when the server makes a connection through it - It will show in your ARP status

 

if you like playing around a bit with Unraid - may i suggest a pfsense box - that's so much more than a wrt!

 

hope this helps!

Thanks everyone.  No matter what I do the MAC addresses on these two servers stay the same.  This still baffles me, but I suppose it doesn't really matter at this point.  The issue is clearly with my router and not with the server itself, so I'll be shipping this server off soon.  I just hope I don't run into this issue again in the future, it is annoying to deal with.  In fact, I also don't understand why I haven't run into this issue in the past...

I've never understood why unRAID allows the hardware mac to be overridden, anyway - how can we stop it recording the mac in the config files?

I've never understood why unRAID allows the hardware mac to be overridden, anyway - how can we stop it recording the mac in the config files?

You send a BUG report to lime-tech.