Guys:

I have been on unraid for a couple years now and never had issues like this. Let me tell you what is going on.

I had a drive that has been installed that started reporting errors during parity check. During this time I started having issues being able to write to shares. Upon reading forums I decided to create users and passwords and regenerated all my mappings. Still I cannot copy data to my server

So I bought one drive to replace the failing one as well as a new replacement parity drive. Now the parity has been recreated and the server is back running with no errors, however I still cannot write data to it.

This morning I tried to create a new share and realized I can't even do that. If I browse my shares on my flash drive, there is a file for my new share but I can't browse to it or see it in the web GUI. I am at my wits end with trying to fix this thing and starting to regret ever building it in the first place. I hope someone can help me fix it to get it back to the worry-free, maintenance-free solution it used to be.

I am using Server Pro 4.7

users

root

admin

brettstid

 

Shares look like this:

name: movies

allocation: high-water

split level: 1

export smb: export read/write

valid users: brettstid

 

Settings

Tower

workgroup: workgroup

local master: no

spin down : 1 hr

force ncq disabled: yes

enable spinup groups: yes

default partition format: MBR unaligned

stripes: 1280

limit: 768

window: 288

 

 

See here for how to attach a syslog for analysis:

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

 

Odds are you have a file system that has been marked as read-only because of detected corruption.

Here is where it gets confusing. Looking at the log there are lines that say both 'Games' which is the share I am trying to create and 'Movies' are both read only and lines that say chmod 700 which should allow read, write, and execute. Any help is appreciated.

 

PS - I did delete the first third due to size restriction

syslog.txt

Actually taking a look at the sys log check this file instead. It may show more of why this message is appearing. There was some info I deleted at the top of the log due to size so I have provided those lines in this file.

syslog2.txt

OK. So I ran across this file system check: http://lime-technology.com/wiki/index.php?title=Check_Disk_Filesystems

Upon running it I discovered that there is an error in disk5, the one I just installed. So, maybe the preclear didn't work or something. Here is the error report at completion:

55: The level of the node (36508) is not correct, (2) expected

the problem in the internal node occured (167102455)finished

Comparing bitmaps..vpf-10640: The on-disk and the correct bitmaps differs.

Bad nodes were found, Semantic pass skipped

20 found corruptions can be fixed only when running with --rebuild-tree

I was told to post here before running that command. What can I do? This is a nearly full 2TB hard drive that will take forever to rebuild.

Run with rebuild-tree.

OK So I replaced the disk with another disk. It rebuild  over the  last 40 hours with no errors. I was able to write a file to the server. This evening I got home and went to write files and got no permission again. So I deleted all users and set settings to simple sharing. I was then able to write some files. About half way through i got a popup out of space. This is due to split level 2 and the drive it was writing to I belive (if I understand split level). So I switched gears and moved to another file, now I get the no permission error again. These servers and the software seems very unstable and so far I haven't found anyone to explain to me why these issues are even occuring. One last note is that now I have 233 errors on the brand new drive I just put in.

OK So I replaced the disk with another disk. It rebuild  over the  last 40 hours with no errors. I was able to write a file to the server. This evening I got home and went to write files and got no permission again. So I deleted all users and set settings to simple sharing. I was then able to write some files. About half way through i got a popup out of space. This is due to split level 2 and the drive it was writing to I belive (if I understand split level). So I switched gears and moved to another file, now I get the no permission error again. These servers and the software seems very unstable and so far I haven't found anyone to explain to me why these issues are even occuring. One last note is that now I have 233 errors on the brand new drive I just put in.

If you replaced a disk with a corrupted file system, unRAId would re-construct the exact same corrupted file system on the replacement disk.  You need to run

reiserfsck --rebuild-tree /dev/md5

as it instructed.    There was probably nothing wrong with the original disk.,

 

The 233 errors can easily be a defective sector on a disk.  The know for sure you would need to post the SMART report from that disk AND also the syslog.  From past experience about 1 in 5 new drives is defective.  You might be the lucky one.

 

Joe L.

OK. So I have been out of town recently, but I have an update. I ran the rebuild tree command and now it shows the drive as unformatted. Now what?

OK. So I have been out of town recently, but I have an update. I ran the rebuild tree command and now it shows the drive as unformatted. Now what?

What was the result of the --rebuild-tree ?  Did you then perform a

reiserfsck --check /dev/mdX

 

In any case, if the --rebuild-tree was successful, reboot the server and the disk should mount.  (unformatted is just what unRAID says whenever a disk cannot be mounted, so it might be formatted, but not mountable because of a failed file system check.)

 

 

I was told to post here before running that command. What can I do? This is a nearly full 2TB hard drive that will take forever to rebuild.

 

Why would you post this and then replace the drive instead of waiting for help?

 

Sometimes questions get lost so you could have bumped the thread or started another one.

 

I'm thinking your newest replacement drive is in worse shape than the previous one since it's showing errors. Follow Joe's advice but it might need to be replaced yet again at some point.

 

Sorry if I jumped the gun. 6 days went by with no response. This server is causing me major problems as my computer is out of space and this is my means of storage so it is hindering me greatly for months. I just want to get it fixed.

At any rate I did reboot. In fact while I was out of town the server was off for 3 days before I turned it on and got the unformatted status. I did run the --check and the results were that the --rebuild-tree did not complete for some reason so I am running it again now. I will repost tomorrow if it finishes by then.

 

Thanks for all your help.

 

CORRECTION - Server is saying this process will take 12 days to complete on a 2TB drive. I will post back then.

The drive rebuild will not likely complete. Post a current syslog.

Actually it did complete, I don't know where my calculation of time went wrong. It deleted and renamed some things. After a reboot it mounted and is now performing a parity check. The parity check normally takes about 30 hours or so, I will post back if all is well on Monday. This is looking up.

OK. So I have been away from the server for some time until today. For some reason every time I turn the thing on it does a parity check for like 36 hours. It started yesterday and today I went to watch a show from it and could not connect to it or get a prompt from the moitor connected directly. I forced a shutdown with the power button (Which I hate to do). There is obviously a problem. When it came back up the parity check started over. I was able to watch some shows while it was checking. I then tried to copy content totalling 50GB to it and it ran for about 20 minutes then failed. Now I cannot get to it via network or through the monitor connected directly. It has power to it but I am unable to view it at all. Any ideas?

A parity check will commence on every reboot following an unclean shutdown. The system should not take 30 hours to do a parity check. It will take equally long to rebuild a failed disk. This is too long to be practical. If the contents of the server is important you should upgrade some of the servers components. 8 hours is a good target.

An unclean power down will cause a parity check.

 

You have to figure out some way to connect to the server and pull a syslog.

 

You could pull the flash and do a check disk on a PC. If you get connected you could also pull SMART reports from the disks and do a reiserfsck on each disk (except parity).

So I know it has been a while, but today I took the time to run the tests. Here are the smart and syslog files. I also ran the resierfsck again and all drives now pass. However when looking at the tower web interface disk 5 shows almost 2,500 errors. Also I know that the system should perform parity check booting from an improper shutdown but mine does after any shutdown. If I issue the "shutdown -h now" command and then an hour later power it back up it will perform a full parity check. Could this be a issue related to the script for scheduling parity check at the first of every month. I don't believe I have this issue after using it, but I can't be for sure. Maybe it's best to fix this hard drive issue first and see if that clears up my other issues.

smart.zip

syslog.zip

Disk 3 needs to be rebuilt. If you have a spare use it now. The drive has a lot of unreadable sectors. None have been reallocated and the disk may still work.

 

shutdown -h is not the correct procedure to shutdown the server. This is an un-clean power down; you may as well be pulling the power cable from the wall. Use the webGUI to shutdown the server.

 

Un-assign disk 3 and assign your spare, if available. Run the pre-clear script on disk, WD-WMAY00197878. The disk is only useful if the Current_Pending_Sector RAW_VALUE goes to zero and stays there. At least 2 pre-clear cycles should be used.  In order to assign WD-WMAY00197878 as disk, 3 after it has been cleared and has a Current_Pending_Sector RAW_VALUE of zero, start the array with disk 3 unassigned then stop the array and assign disk 3.

Wait. Disk 3 has the problems? WMAY00197878 is my disk 5 that is showing errors. Are you sure you are reading this right? That disk 5 is nearly brand new and is a FULL 2TB with data. I will await further advice.

My mistake. Disk 5 it is.