My unRaid tower has been getting worse over the last month. It will stop responding while doing any sort of simple task (creating a folder, reading a file). So I restart the tower, it does a parity check and then I wait until it does it again. It started out doing it once a week or so and now it is down to even doing it during the parity check. My tower is essentially useless at this point because doing anything will cause it to freeze and become unresponsive.

 

Does anyone have a clue what is going on here and how it can be fixed?

The syslog looks like the syslog after a fresh reboot.  I don't see anything wrong in there, but hopefully RobJ or Joe L. will take a look and give a more authoritative review.

 

There are two very common causes of unRAID crashes:

1-Bad memory

2-Disk cabling / backplane issues

 

Bad memory is easy to test by selecting the memory test from the boot menu.

 

Cabling problems can either be with the data cable or with the power cable.  If there is any loss of connection, even a pretty short one, the computer can lose connection and the dirve.  This does not always cause a crash, but frequently does.

 

You might want to open up the case and double check all of the disk connections.  If that doesn't work, I'll tell you from experience that these types of problems can be hard to track down.  I had a problem that took hours of isolating and testing.  As a part I replaced some of my SATA cables with the locking variety.  I recommend locking cables where they can be used.

 

One idea for being able to see the last few log entries (which might be helpful), is to run this command at the command prompt on the unRAID server.

 

tail -f /var/log/syslog

 

When the server crashes, hopefully the last screenful on error messages will be present.  You can use a digital camera to take a picture and post.

Like Brian, I don't see anything in the syslog either.

 

I lean more to a memory or other failing hardware component.  Cable and connection issues are almost always caught by the system, and reported as errors, they don't usually cause crashes or hangs.  A memory test and tailing the syslog are very good ideas.  It sounds like these crashes are fairly random, but *may* be occurring more often when doing something, rather than at random idle moments?  Then one other possibility is a failing power supply.

If it can't handle a simple task, then I also think about a hardware problem like memory or PSU.

Did you add some hardware before it started to freeze ?

If so, there could be the problem. If not, test the memory and/or try with another PSU.

 

Thank you guys. The memory test came back with 0 errors. I have not added any hardware but recently cleaned everything up by removing everything from the case and putting it back together again. It does seem like it could be a power supply problem so I have ordered the Corsair CMPSU-650TX (I have wanted to upgrade to it for a while anyways).

 

I tried using the line:

 

tail -f /var/log/syslog

 

but did not receive any error message when the server crashed again.

 

If there are anymore ideas I'd be happy to try them but I will definitely post back when I receive the new power supply.

I ended up replacing everything since the machine was 3 years old and seemed to be on its last leg. Here's the new setup:

 

GIGABYTE GA-MA74GM-S2

Kingston 2GB (2 x 1GB)

CORSAIR CMPSU-650TX Power Supply

AMD|A64 X2 5200

 

I got everything put back in the case and I started to set my drives in the correct position, then the machine stopped. I put the Flash Drive in my desktop and it what I had was 1 very strangly named file in the config folder. All other files were gone, even the key file. So I formatted the USB Drive again, installed 4.5beta6 and put my key back on the flash drive. I got to 18% and it stopped responding again.

 

Now I cannot startup the tower with the flash drive connected. At this point I have to think the flash drive is going bad. Any suggestions?

 

I have replaced the memory and power supply in the tower but the problem still persisted. I turned off user shares and it seems to be fixed.

 

I thought maybe it was a problem with 4.4.2 so I upgraded to 4.5beta6 and the problem is still there when user shares are turned on. I have not found a fix yet, but I thought I would post on what I found out.

 

The problem is more then likely not with 4.4.2 as that is what I am running along with a quite a few on this forum.

 

Your problem sounds like a hardware problem.  You might be better off to just take everything out of the case and restart the build.  If you could list the components you are using this might help.  Also, check to make sure that the RAM timings are set correctly.  Some of the time the motherboard can get it wrong, so its good to do some checking and setting for yourself.

prostuff, I went ahead and edited the post above with more information before I noticed you had responded.

Try the free version of unRAID with three disks with a different flash drive.  If you can get it to work reliably then it is more then likely that your flash is starting to take a crap.

aitf311,

 

Have you noticed the the lockup occurs during a file transfer?  If so, can you repeat the lockup by transferring the same file again?

 

I was experiencing very similar symptoms.  I thought I had replaced absolutely every piece of hardware and I still had the problem.  See:

 

http://lime-technology.com/forum/index.php?topic=3000.0

 

Still, the problem wasn't going away but it was tollerable (I didn't have a choice but to deal with periodic lockups since I don't have any other machine with 7TB of storage).  Then some time later while trying to load several large ISOs I got frustrated enough to ask for help again:

 

http://lime-technology.com/forum/index.php?topic=3809.0

 

Where I am now...

 

What I noticed most recently was that there were 4 particular ISOs that would not copy without locking up my server.  As strange as this (still) sounds to me it is a fact that I couldn't deny as the failure was repeatable.

 

It occured to me that I was incorrect about having replaced 100% of the hardware.  There was still the network card. 

 

I replaced the network card and was then able to copy the 4 troublesome ISOs.  My server hasn't locked up for almost a month now.  The jury is still out, but I have my fingers crossed.

 

jarDo

 

 

JarDo,

 

The problem is definitely not limited to transfers only. It crashes while trying to read, making a new folder and even doing a parity check. It seems to be even easier to trigger on user shares and especially my main media folder than contains links to my main user shares.

 

Prostuff,

 

If I do that with the first 3 disks will I have any problem with the other 4 disks I have in the array when I go back to my pro key?

 

 

I just restarted my machine after a successful parity check just to see if there would be any problems and sure enough the parity drive is reporting as invalid. I did a full parity check last night and everything came back as good.

After rebooting with the bad parity this morning I let the parity run a full time through again. It went without a problem. I then tried to move a file over from my computer and the tower halted again.

After rebooting with the bad parity this morning I let the parity run a full time through again. It went without a problem. I then tried to move a file over from my computer and the tower halted again.

 

I had a problem similar to your once i set everything up and it turned out one of NIC's on my Abit AB9 Pro was screwed up/bad.  If you have an PCI NIC card try putting it in the board to see if that fixes it.  Since my board has 2 NIC's i disabled the bad one in BIOS and then just used the other one.

OK I will try that this evening but I did replace the motherboard and I am using the on-board nic with the same problems as before.

OK I will try that this evening but I did replace the motherboard and I am using the on-board nic with the same problems as before.

 

That's what happened to me.  I was using the onboard NIC prior to, and after, replacing the MB.  In my mind I was running a "new" NIC.  But it wasn't until I switched to a PCI NIC (that stated on the box was LINUX comaptible) did my problems seem to go away.

OK I will try that this evening but I did replace the motherboard and I am using the on-board nic with the same problems as before.

 

That's what happened to me.  I was using the onboard NIC prior to, and after, replacing the MB.  In my mind I was running a "new" NIC.  But it wasn't until I switched to a PCI NIC (that stated on the box was LINUX comaptible) did my problems seem to go away.

It has not been made clear in your posts... When your server "dies" can you still log in via the system console? or not?

 

I know you tried a "tail -f /var/log/syslog"

Did you try that from a telnet window on your PC, or from the system console?  If you did not do it on the system console, then that is something else to try.

 

Only reason I'm asking is seldom does Linux just stop without errors.  If the "tail -f" was done from your PC, via telnet, then it might be the network link that is dropping when the server crashes so you do not get to see the actual error.  If you did the "tail -f" from the system console, you might get to see the error.

 

Joe L.

Your last syslog looks OK except for one hiccup at the end:

Jun 18 11:24:35 Tower kernel: usb 1-4: reset high speed USB device using ehci_hcd and address 2

That appears to be at the address of your USB flash drive.  There are no errors reported, so I don't want to conclude anything here, but a reset like that is highly unusual, would seem to indicate something odd about that USB connection or the drive itself.  That coupled with your following statement, makes your flash drive a little suspicious.

I just restarted my machine after a successful parity check just to see if there would be any problems and sure enough the parity drive is reporting as invalid. I did a full parity check last night and everything came back as good.

That behavior is usually associated with a crash or bad shutdown, or a bad flash, or lost contact with the flash before the previous shutdown.  However you may have several problems, because a bad flash is unlikely to cause any of the crashes you have been experiencing.

 

It's a little confusing to me what is your current situation, because you have changed out so many things.  Can you provide more detail of what goes wrong now, and what errors you see now?  Joe's instructions are very good.  By the way, have you tried a memory test of the new memory?

Joe, I did try it from telnet last time. I just tried the tail command from the system command and have posted the screenshot. It is going through a parity check right now.

 

When I say die what I mean is that I can no longer access the server via smb shares. The web interface is completely broken except the refresh button during a parity check. I can see the percentage continue to go up, but even after it finishes I cannot access the drives.

 

I *think* I can access the tower via telnet and system command when the shares and web interface stop responding, but I will make sure after the parity check is finished and it happens again.

 

With the new build I am having to continually reformat the flash drive because somehow it becomes corrupt during this process and the system will not boot up until I do.

 

Rob, I will make sure to test the new memory once the parity finishes up.

With the new build I am having to continually reformat the flash drive because somehow it becomes corrupt during this process and the system will not boot up until I do.

I'd call that a third strike against this flash drive.  It seems very unlikely that both the old and new motherboards could be corrupting the flash.  You are making sure that when you prepare this flash drive on a Windows machine, to use the "Safely Remove Hardware" icon before removing it?

 

See the FAQ entry, "If my flash drive dies or is lost, will I lose my license?"

Yes, I am flashing on a windows machine and safely ejecting it. I am using syslinux -ma to prepare the drive as well.

After replacing the NIC card yesterday, running a parity check, running a memory check and having all things come back without a problem I tried to test the tower by copying files to and from the tower and going in to my user shares. This is the first time in a very long time I have been able to play with it for 10+ minutes and it not drop the connection.

 

If it keeps up, I have to be convinced that the NIC card was in fact the problem but I will report back tonight after a full day of up-time!

Recently I had an issue I couldn't figure out w/random freezing (not w/unraid but in w2k3 server) and after hours spent troubleshooting I replaced the power supply on a whim and all is well.  You mentioned some new components, but not a new PS - it's a long shot, but if it happened to me it has to happen elsewhere.  Hopefully you resolved it w/the NIC card though.

 

Nate

I've had one full day now of using it without any kind of problems. This is the first time it's went this long without a problem for at least a month now so it feels like the problem has been fixed!