Good evening,

First off let me warn you, while I'm no slouch when it comes to computers I am a novice in regards to the nuances of UnRaid. It took me forever to get comfortable with something as mundane as Telnet. So please be kind. Talk SLOW and use SMALL words. 

 

 

I use my UnRaid Server as a Media Server. Mostly to watch movies that I had converted to MKV files via a computer or a WDTV that or to view family photos. About two months ago we noticed some strange behavior while watching movies.

 

 

(I should mention that within the same time frame I had upgraded to UnMenu, and then added a UPC APC & interface cable along with the appropriate plug ins to safely shut the server down in case of a long term power outage. Both of these items have performed flawlessly. I only mention them in case it may be relevant)

 

 

At times it would kick you out. I also notices that the server would deny me when it came time to move or rename a file on the server. Almost as if I lacked the permission. That is weird because I was able to do this with no problems months before when I copied the movies over.

 

 

I started noticing these REISERFS messages popping up on the server console.

I took a photo of the screen (UnRaid Server Message1.jpg)

 

 

I looked up this problem in the FAQ  and found the proper steps to take here:

 

 

http://lime-technology.com/wiki/index.php/Check_Disk_Filesystems#Preparing_to_run_reiserfsck

 

 

I ran a Memtest with no errors. Then from the UnMenu "Disk management" tab I ran a Long Smart Test (smart test report.txt)

 

 

from the UnMenu "Disk management" tab again I ran a FileCheck (which I assume is nothing more than reiserfsck --check /dev/md1)

 

 

After reading the results of this I ran

reiserfsck --fix--fixable /dev/md1

 

 

THIS showed me (what looks like a large amount of) errors on the hard drive.

 

 

I saw something in the report that mentioned running reiserfsh with the ---rebuild-tree option. But from what I read, it's not a thing to be done lightly.

 

 

So I've uploaded as much info as I thought would be helpful. Can anyone help me out here? I need to know what the next step is? Is this hard drive bad? It's pretty new. I got it from NewEgg not 4 months ago.

 

 

 

 

Just wanted to add the system specs.

• Motherboard is a Biostar M7VIT 800 with a AMD 1Ghz CPU
  
• RAM: 1 GB
  
• SATA Controller: 4 Port SATA Serial ATA PCI RAID Controller Card - Silicon Image
  
• NIC: Intellinet Model 522328 100/1000MBps
  
• Hard Drives: (2) Western Digital WD Green WD20EZRX 2TB IntelliPower 64MB Cache SATA 6.0Gb/s 3.5" Internal Hard Drive Bare Drive (One as Parity and one for Data) & (1) Seagate 250 GB IDE
  

here are the rest of the files. The Syslog had to be zipped as did the smart test report

filecheck.zip

syslog-2014-08-13.zip

smart_test_report.txt

Just remembered something I did before this that may be part of the problem (or not) 

 

 

Just before this happened I added some User Shares. Basically I just mapped the directory //tower/disk1/video to "Video" and  //tower/disk1/Photos to "Photos"

 

 

And is there a way to GET RID of these user shares? There should be, right?

 

 

Edit to add: Was reading through the UnRaid manual and came across this tidbit (so never mind that last qusetion!)

 

 

Deleting User Shares

To delete a User Share, just clear the Share name field and click Apply. Only entirely empty User Shares may be deleted.

Anyone?

Just remembered something I did before this that may be part of the problem (or not) 

 

 

Just before this happened I added some User Shares. Basically I just mapped the directory //tower/disk1/video to "Video" and  //tower/disk1/Photos to "Photos"

 

 

And is there a way to GET RID of these user shares? There should be, right?

 

 

Edit to add: Was reading through the UnRaid manual and came across this tidbit (so never mind that last qusetion!)

 

 

Deleting User Shares

To delete a User Share, just clear the Share name field and click Apply. Only entirely empty User Shares may be deleted.

I'll let someone with more experience help you with reiserfsck, but I can comment on your user shares. Any folder at the root of cache or an array disk is automatically a user share. If you do not configure a share, it has default configuration.

 

Not sure what you mean by the word "mapped" here. If you mean creating or configuring a user share, you should be aware that Linux is case-sensitive, so a folder named video and a folder named Video are not the same folder.

Run with rebuild-tree.

It appears you had a hard crash. The syslog also shows that your flash is corrupted. I am not sure what happened.

 

From your photograph, I believe that there are issues with the OS (or more specifically the Reiser file system driver) talking to the drive. This has resulted in disk corruption.

 

How long have you been running unRAID? Are you certain the MB is 100% compatible? It is pretty old. Not saying it won't work, but with some of the exciting new features in unRAID, you might consider an upgrade.

 

I am not sure exactly what caused the corruption. But based on the check, it is corrupted. The ONLY way to fix the corruption is to run with the suggested option. But if your computer is having issues talking to the drives, you might want to hold off running it until you have either solved the mystery or changed the system into something that does not show these symptoms.

 

Maybe someone else would be of more help.

It appears you had a hard crash. The syslog also shows that your flash is corrupted. I am not sure what happened.

 

The flash is corrupted? That's the second one I've gone through. The first was a crappy flash drive. I could see how that would get corrupted. It was a piece of junk. This second Flash Drive is a Kingdston 8Gb. It's sort of old but was never really used. I switched to a 32 Gig soon after I bought it. So it sat unused.

How long have you been running unRAID? Are you certain the MB is 100% compatible? It is pretty old. Not saying it won't work, but with some of the exciting new features in unRAID, you might consider an upgrade.

 

 

I've been running UnRaid since January. I started off with some old IDE Drives and then

about 5 months ago I upgraded to (2) 2TB SATA Drives and New SATA controller.

 

 

How would I find if the MB is compatible or not?

 

 

I am not sure exactly what caused the corruption. But based on the check, it is corrupted. The ONLY way to fix the corruption is to run with the suggested option. But if your computer is having issues talking to the drives, you might want to hold off running it until you have either solved the mystery or changed the system into something that does not show these symptoms.

 

 

When you say "run with the suggested option" you mean "--rebuild-tree"?

A corrupted flash does not necessarily mean the flash has failed. It usually just means that the file system needs repair. This often caused by an unsafe shutdown, similar to pulling it out without safely removing it in Windows. Put it in your PC and run checkdisk.

Just safely shut down the UnRaid server and scanned the flash drive.

 

 

No problems were noted

I looked into the MB Compatibility (using the list http://lime-technology.com/wiki/index.php/Hardware_Compatibility#Motherboard) issue and wasn't able to get a determination. It's not one of the MBs listed.

 

I am using a Biostar M7VIT 800 V7.X

http://www.biostar-usa.com/mbdetails.asp?model=m7vit+800+v7.x

 

 

Yeah, it's old. But it was working for at least 4 months. Wouldn't a Memtest show if there was some kind of issue with the MB?

 

 

While this MB has an on board SATA Controller, I opted to get a separate one. The SATA controller I'm using is one that is compatible with UnRaid

 

 

SATA Controller: 4 Port SATA Serial ATA PCI RAID Controller Card - Silicon Image

http://www.monoprice.com/Product?c_id=104&cp_id=10407&cs_id=1040702&p_id=2667&seq=1&format=2

 

 

For now I've taken the server down and it sits unpowered.

 

 

What's the next step? I'm assuming it running  reiserfsck --rebuild-tree /dev/md1

 

 

 

 

I looked into the MB Compatibility (using the list http://lime-technology.com/wiki/index.php/Hardware_Compatibility#Motherboard) issue and wasn't able to get a determination. It's not one of the MBs listed.

 

I am using a Biostar M7VIT 800 V7.X

http://www.biostar-usa.com/mbdetails.asp?model=m7vit+800+v7.x

 

 

Yeah, it's old. But it was working for at least 4 months. Wouldn't a Memtest show if there was some kind of issue with the MB?

 

 

While this MB has an on board SATA Controller, I opted to get a separate one. The SATA controller I'm using is one that is compatible with UnRaid

 

 

SATA Controller: 4 Port SATA Serial ATA PCI RAID Controller Card - Silicon Image

http://www.monoprice.com/Product?c_id=104&cp_id=10407&cs_id=1040702&p_id=2667&seq=1&format=2

 

 

For now I've taken the server down and it sits unpowered.

 

 

What's the next step? I'm assuming it running  reiserfsck --rebuild-tree /dev/md1

 

If the server has been running fine for 4 months, and you have used it pretty heavily (copying lots of data to it, reading data from it, performing a few parity checks) that compatibility is likely not the issue.

 

Memtest is a good idea to confirm that your RAM is solid. Bad RAM could cause such a crash.

 

Addons/plugins. Some addons can interfere with unRAID. If you have any loaded you might consider running in "safe mode" (it is a boot option). If you added something new recently, that would be a prime suspect.

 

Drive connections. Data and/or power cabling problems can show up in various ways. Sometimes a disk is red balled, but sometimes it can cause resets and other problems. Also, although rare, sometimes the way that sata cables are run, and if they are tightly bundled, can cause crosstalk and interfere with drive signals.

 

But if you feel all i soundly connected and that the server is healthy hardware wise, then yes I would think you could run reiserfsck as the check option suggested.

Thanks for the tips. I like a methodological approach. Now I have a good idea where to proceed from here.

 

 

Since I already ran Memtest and didn't get a single error, I'm thinking that's a very good thing. 

 

 

 

I'll check the HD connections first and make sure nothing is loose or cables within too close of a proximity.

 

 

After that I'll boot the server in Safe mode and see how she behaves. I'll report back with my results.

 

 

Thanks again.

With the server safely powered down I checked all the HD SATA cables and their connectors. I even went as far as the re-seat NIC, Video, and SATA cards.

 

 

I then powered it up in SAFE mode and allowed it to run for about 3 hours. Within 60 minutes I streamed a video from this server to my phone, then to my computer, and then my WDTV. Watched them all with no stutters skips or problems.

 

 

After this I checked the server console. There were no error messages shown. So I was thinking "Good!"

 

 

So I went and took care of some chores and checked back at the console. And yes, there were a bunch of REISERFS errors.  >

 

 

What's odd is that NOTHING on this server was being accessed. It's pretty much just sitting there.

 

 

So I guess it's time for the rebuild tree option

 

Ok, I ran reiserfsck --rebuild-tree /dev/md1

 

 

After that I ran a parity check and left for the day. I came home today and the console was filled up with REISERFS Errors  :'( :'( :'(

 

 

What now?

 

 

 

 

Anyone?

While I was working to resolve this issue I remembered that this computer has an IDE drive along with the (2) SATA drives.

 

 

I thought to myself, "Could this be the problem?" Since the IDE drive was only 250G I decided to remove it and see if that cleared up the REISERFS issues.

 

 

So I followed the directions from the FAQ and removed the drive first from UnRaid. Then last night I rebooted the server and allowed it to rebuild the Parity. Today I came home and the Parity check was complete. So I shut the server down and physically removed the IDE drive. After that I powered up and allowed it to start.

 

 

I allowed the server to run for about 2 hours without accessing anything on it. Then I tried watching a movie. It SEEMED to work fine.

 

 

Then I checked the console.

 

 

REISERFS errors all over again

 

 

Could it be because there are two SATA controllers? I have one on the MB (not connected to anything) and one in a PCI slot.

 

 

What is my next step?

 

 

 

 

 

 

 

 

See check disk filesystems in my sig.

I took your advice even thought I had run REISERFSCK --fix-fixable and REISERFSCK --rebuild-tree before.

 

 

I figured that it could not hurt and that I had changed the set up a bit by eliminating that old IDE drive from the mix.

 

 

So last night I ran REISERFSCK --fix-fixable it generated a notice about the need to run REISERFSCK --rebuild-tree.

 

At first there was a little action and I thought the server had locked up. But I let it go and after several minutes the display changed and work began. The program stated it had about 11104 seconds left to run which I calculated as just over 3 hours. So I let it go and checked it this morning. It appeared that the program had finished and I was back to the normal command prompt. So I safely shut the server down and then started it again. when i get home tonight I will post an update.

 

 

 

 

 

 

 

 

 

Got home last night and immediately looked at the server console. NO ERROR MESSAGES!

 

 

So I queued up some movies and watched them using several different sources. There were no problems or errors.

 

 

I will keep an eye on the server and see if any further problems crop up. Until then I'm going to marked this as SOLVED