Newbie to UnRaid, experienced in both building systems and servers (WHS, Database, etc).  Adequate knowledge of Linux.

 

Fresh build (all repurposed equipment)

Asus P8Z68-V Pro i5 2500K, 16GB Memory

Supermicro AOC-SAS2LPMV8 8 Port HBA

Syba 2 Port SSD Caddy (for Cache and Apps)

 

Drives:

Parity: 4TB Seagate

Cache: 750GB Samsung EVO

Apps: 128GB Samsung Pro

Data1: 3TB Hitachi (on Mobo)

Data2: 3TB Hitachi (on Mobo)

Data3: 2TB Hitachi (on AOC)

Data4: 2TB Hitachi (AOC)

Data5: 2TB Hitachi (AOC)

Data6: 2TB Hitachi (AOC)

Data7: 2TB Hitachi (AOC)

Data8: 2TB Hitachi (AOC)

Data9: 2TB Hitachi (AOC)

Data10: 2TB Hitachi (AOC)

 

USB 3.0 32GB SanDisk

UnRaid version: unRAIDServer-6.0-beta14b-x86_64.zip

 

Follow procedure to copy files to USB and make bootable.

 

First boot (all defaults, no edits to any file), the server comes up just fine, all files load, no errors, Tower login displays.

 

HTTP://Tower display from web page and I can see the server and

can access the server web page.

 

I started to configure Parity and 2 drives (I only have the demo for the time being to make sure I can get this working).

 

Then the server panics.  The server stays running for about 2 minutes, then I get a Kernal Panic.  There is a trace, then kernel panic - not syncing: fatal exception in interrupt.

 

Kernel offset: 0x0 0xffffffff81000000

 

I prepped the drives by deleting the partitions but did not pre-clear the drives.

 

Is this a pre-clear problem or something else?

 

Thanks in advance for your wisdom...

Jeff...

Number one step by far is to thoroughly test your memory.  Memtest is available on the boot menu, and I would let it run all night (unless it fails earlier).

Next possibility would be to switch the thumb drive to a USB 2.0 socket, as 3.0 is still a little buggy with some boards.

Preclearing the drives is not necessary in such a scenario, but would not be a bad idea.  If nothing else it will test the drives, disk controllers and cabling used for the preclear.  I have certainly seen bad drives cause strange errors.

I even had a bad drive one time lockup unRAID when I was attempting to run the preclear script on it!  It took a hard reboot to recover! 

 

Note for the OP:  Here is a link to the preclear script: 

 

        http://lime-technology.com/forum/index.php?topic=2817.0

 

By the way, you can run more than instance of preclear at a time using either screen (if using telnet) or the console (You use alt-F1 through alt-F6 to switch sessions).  Most experienced users run at least one (some who have had the tails bitten, run three!) cycle.  It tests the drive before adding it to the array.  Plus it puts enough hours on the drive to get it by the infant morality point.  (Google 'bathtub curve' for details.  Back in the dark ages of computers, every vendor burned a computer in for 72 hours before delivering it to the customer!) )

Hey Thanks all for your posts.  I ran the memtest all night.  No issues.  Then I followed JonathanM's advice and moved from 3.0 to 2.0 socket.  It worked longer.  Ran for about 30 minutes before it panicked again.  Any other ideas? 

 

Here's a quick update. 

 

Has anyone ever run into a situation where a panicked UnRaid server, when it is running, messes with your network?  I first noticed it because it seemed like my internet had dropped.  Even called my hosting company before I figured out that the network goes out when the UnRaid server has crashed!  I can predictably turn the problem on and off - just start the UnRaid server, let it run a couple of minutes til it panics, then boom, my internet goes out.  Turn off the unraid server and instantly the network is back.  This diagnostic is predictable:

 

1. My PC is playing video (news program)

2. Unraid server starts.

3. Video continues to run.

4. Unraid (I've done nothing on the server, just booted it) after 3-4 minutes it panics (same info as above).

5. My PC video hangs.  No network, can't access any local or online resource.

6. Turn Unraid server off.

7.  Immediately, all network resources work (no restarts, or anything).

 

Somehow this feels like the panicked UnRaid server is flooding the network?  Is that diagnostic as to why its crashing?

 

This same machine can run for days (exact configuration) running win 7 off the SSD.

 

Thoughts?

Ok, I've tried to take this down to the bare minimum.

 

1. I removed everything from the system except what is minimally required.

- Mobo, memory, no cards, only using MOBO Sata.  3 drives.

2. Rebooted and server starts as expected

3. Loaded Tower web page

4. Started to config and within 60 seconds the system panics.

 

Oh and just to confirm, with the server in Panic mode, my network is down!  Until I turn the server off.  Then everything works fine.

 

This exact hardware runs for days on Windows 7 with no issues what-so-ever.  I've run 24 hour memory test, no issues.

 

I don't know what else to do.

 

At this point, my only option is to move on to another platform.  I've had WHS working on this hardware before, I suppose I could try some others.  Anyone have any wisdom?  I really want to make UnRaid work - I love the technology concept, but I'm stuck at this point...frustrating...

Read the first post in this thread:

 

      http://lime-technology.com/forum/index.php?topic=9880.0

 

You find to run 'tail' command in a telnet session and post the output here after the freeze happens.  Hopefully, something will show up with that logging of the state of the computer at the time when it locks up.

 

 

Does it panic if you disconnect the network?

That's a great question.  I ran the system without the ethernet cable attached.  Ran for a long time without crashing.

 

I also ran the latest 5x version.  That ran for many hours without crashing either. 

 

So it sounds connected to the ethernet somehow?  Driver problem?  I can access HTTP://Tower just fine until it crashes...

Read the first post in this thread:

 

      http://lime-technology.com/forum/index.php?topic=9880.0

 

You find to run 'tail' command in a telnet session and post the output here after the freeze happens.  Hopefully, something will show up with that logging of the state of the computer at the time when it locks up.

 

Frank, I'd love to do that, but unfortunately, when it crashes, the system is locked and the network flooded (i.e. I can't access anything on the local network).  Is there a log archived somewhere?

Frank, I just ran another test, since i can't get at a log, here's the closest thing I can get to a tail of the log.  This is what displays on the screen.

 

http://my.jetscreenshot.com/12412/20150410-wee7-1655kb

 

All the info regarding my config, versions, etc is in my first post.  Let me know if you need more...

 

Thanks in advance...

I did a search on your board and the NIC on it (Intel® 82579). Someone using Ubuntu had a weird driver issue. Might be worth a read - No idea if it will help anyone at all but thought id post it.

 

http://ubuntuforums.org/showthread.php?t=2167864

Frank, I just ran another test, since i can't get at a log, here's the closest thing I can get to a tail of the log.  This is what displays on the screen.

 

http://my.jetscreenshot.com/12412/20150410-wee7-1655kb

 

All the info regarding my config, versions, etc is in my first post.  Let me know if you need more...

 

Thanks in advance...

 

Does not look good.  I am not much of an expert on such manners but I do have an investigative mind.  I assume that you were connected to your network with a cat 5 cable at this point...  correct?  If you are not having a problem with the cat5 cable unconnected, I would also be suspecting a driver problem with the NIC.  In that case, LimeTech needs to be involved...  (BTW, NIC problems have been a sore spot in the past...)

 

 

I followed your link on the unbuntu find.  Good catch.  However, deep in the discussion there was this point:

 

"I suggest you retrace your steps and compile at least 2.3.2 which works well on my 13.04 system."

 

I just did a modinfo on e1000e, my version is 2.5.4.  So sounds like that solution isn't applicable?

 

So I'm convinced now that I've surfaced a bug, especially since the same hardware works fine under the latest UnRaid v5.0.6.

 

How do I connect with Lime Tech to report the problem and hopefully get a fix?

On the MAin page for this forum, you will see a Topic heading "Defect Reports" right above the Topic heading (General Support") that you posted this Thread in.

Defect reported: http://lime-technology.com/forum/index.php?topic=39215.0

 

Thanks for everyone's help...

 

Jeff...

That's a great question.  I ran the system without the ethernet cable attached.  Ran for a long time without crashing.

 

I also ran the latest 5x version.  That ran for many hours without crashing either. 

 

So it sounds connected to the ethernet somehow?  Driver problem?  I can access HTTP://Tower just fine until it crashes...

 

Do you have this system set up for dual-boot by any chance?  I had a somewhat similar problem where if I was running Windows, then rebooted to Linux, the network interface would work for about 5 minutes, then die.

 

I had power-cycled the system from the  "soft" power switch on the front on the machine, with no change in behavior.  I unplugged the system from AC power and let it sit for 5 minutes.  After that I booted straight into Linux, and the system has been fine for a year, including power cycling.  I just can't boot to Windows and back to Linux without disconnecting from AC power.

 

 

Interesting question John, in its original configuration, the system was set to boot to Win 7, Win 8 and Server 2012.  That hard drive isn't in this system any more.  So the answer, for the purposes of this issue is no, dual boot is not enabled.  In fact ALL the drives are unformatted at this point.

So the answer, for the purposes of this issue is no, dual boot is not enabled.

 

Hmm...  Along the same lines, do you have another network interface you could try, even if only temporarily?

I do have an old PCI 10Base100 PCI card I could use.  Would I have to disable the port on the mobo first or would unraid figure out the right driver for this old card (Trendnet TE100-PCI).

You shouldn't have to do anything to your mobo - just install the new card and test.  Since you mentioned the server is stable as long as the cable is detached, you shouldn't have any problems there.

 

I've had problems with those Intel 82579 controllers in the past with Linux.  Driver updates were spotty at best, so we just replaced the cards with some other card in all of our servers.

Tried the Trendnet adapter, while the adapter appears to be active (ifconfig shows it has an ip address), I'm not able to ping anything except 127.0.0.1.  Ping of the server from the network gets host unreachable.

 

The port shows active in my network management software and it shows 100 mbit speed (which is correct).  But can't see the server.

 

Open to any ideas...

 

Disable mobo NIC in BIOS.