I am having a problem I need help with.  First I will summarize and then give the backstory.

 

Bottom line:  I have added a 3 tb data drive to my array, which is listed as "empty" in the Unmenu "mymain" dialog.  I am unable to copy any data to this drive.  No errors are listed, but when I copy files to the array, none of the files go to this drive, even though they should go only to this drive.  I am using the "most free" allocation method.

 

Backstory (nothing bad here; I dont think any of this has caused a problem):

 

I recently upgraded my array with two 3 tb drives, and replaced a 1 tb drive with a 2 tb drive.  I began with a 2 tb parity drive and a spare, precleared 2 tb drive.  The array was working fine, and a parity check showed no errors.  I shut down the array and attached one 3tb drive and precleared it with the array running, copying data to it while preclearing.  After finishing, I shut down the array again, connected the cleared 3tb drive as parity, removed the old 2tb parity, and connected the second 3tb drive and my spare 2 tb drive (making the old parity drive my new spare).  I brought up the array and allowed it to rebuild parity, no problems noted.  I then precleared the other drives and shut down the array again - to physically move drives around.  I removed one of my 1tb data drives that was 80 % full. 

 

Then I brought up the array , and replaced the 1tb data drive with the 2tb data drive, and allowed the array to rebuild that drive.  It worked fine.  I then added the other 3 tb drive to the array.  The array did not recognize the drive as cleared and immediately began clearing it.  After that annoyance, I formatted the drive and added it to the array.  Then I checked parity; no errors.

 

Now, my array is working fine, except that all files are being added to the 2 tb drive that I added which is now about 35 % full.  The new 3tb drive remains empty.

 

What could be wrong?  I posted this problem in the 5b13 beta thread as well; here is that link for reference.  There appears to be a problem with the 3tb disk, disk 9, but I am unable to diagnose it. 

 

http://lime-technology.com/forum/index.php?topic=16125.msg168306#msg168306

 

thanks for your help. 

syslog-2012-03-05.txt

Are you going to share the rest of your share settings and the file structure you are using?

There is something wrong with disk9:

mount: wrong fs type, bad option, bad superblock on /dev/md9,
Mar  3 16:54:44 Repository logger:        missing codepage or helper program, or other error
Mar  3 16:54:44 Repository logger:        In some cases useful info is found in syslog - try
Mar  3 16:54:44 Repository logger:        dmesg | tail  or so

The directory structure is pretty simple.  As I said in the OP, I am using the "most free" setting, and a level of 4.  The pertinent Share is called HomeTube - I am adding files in the following share level directory

 

//Hometube/Film/abc.mkv

 

so, the disk level directory looks like this: 

 

/mnt/disk8/HomeTube/Film/abc.mkv

/mnt/disk9/.

 

This directory is on every drive, except the new 3tb.  As my other drives have 400 mb free space on them and disk8 has 1.3 tb on it, all of the new data is going to disk8.  However, disk9 is empty and there is a problem with it, as dgaschk points out. 

 

Thanks for your help.

I would unassign the disk and re-initialize so that parity is build without it again. You can then either preclear it again or use dd and clear the first part of the disk and then add it again.

Thanks, Lionel, I will try that tonight. 

 

Can you explain "use dd and clear the first part of the disk and then add it again"?

 

I dont understand this part.  What is dd?

dd is a linux command. A google search would give you more info. The basic idea is to use it to clear the MBR of the disk. "dd if=/dev/zero of=/dev/sdX bs=1000 count=1" will do this. You will have to figure out what device to use for sdX. Look at the disk assignment, it should appear there with the 3 letters. sdf for example. Don't get it wrong or you will destroy the data on another disk.

 

Peter

 

OK Peter.  Now I get it, I understand the device naming, no problem.  Not sure it's in the syslog but this drive is sdk iirc (I'm not at home).  I'll also google dd and give this a try tonight. 

 

thanks again, and I'll let you know what I find. 

Wait a minute, was the mistake I made that I assigned my new disk _before_ starting the array? 

New drives must be assigned before starting the array. Instead of dd you can use pre-clear. Enter "preclear_disk.sh -?" for help. Are you using the latest version of preclear?

 

I guess I didnt write it well, but yes, this is exactly what I did.  I precleared the drive properly (although I think I forgot to use -A - how critical is this??), stopped the array, assigned it to a new, previously unoccupied slot, and started the array.  The array immediately became unavailable and it appeared the drive went through the write cycle of a preclear, judging by the time it took - about 9 hr.  Then the array was back online, showing the 3 tb disk as part of the array, and empty.  But files do not get copied to it.  There is no evidence of a problem in mymain/unmenu or on the main array page. 

 

I removed the drive from the array and ran initconfig.  I am now building parity again. 

btw, I am running preclear version 1.12.  I will upgrade to 1.13.

OK, so my array has rebuilt parity, and I am running a parity check since theres nothing to do during the day.  What should I do next?  MY new disk still needs to be added to the array.  I am pretty sure the array remembers the disk as having been cleared because the large number of reads and writes remains listed in the display on mymain/unmenu.  I don't want to preclear the drive again and then clear it again - that is too many write cycles. 

 

Will the dd command help or hurt in this regard? 

 

Do I follow the normal procedure of stopping the array, adding the disk to an empty slot, and then restarting the array?  Or is some other procedure required?

 

I am not sure what to do.  Please advise. 

You either have to preclear or use dd and let unRAID clear it. The disk should add without unRAID clearing it after using the preclear script. It might not have worked  because of the older script version and the 3T drive. I don't know the preclear script history but I believe it had to be modified to properly "tag" a drive >2T after it had been cleared. The newest version might do this properly while the older version did not.

 

Peter

 

I don't know the preclear script history but I believe it had to be modified to properly "tag" a drive >2T after it had been cleared. The newest version might do this properly while the older version did not.

 

Peter

correct.  Older versions of the preclear script did not know how to set the signature on a 3TB drive.  (they pre-cleared it though) Therefore, it was not recognized as cleared.

 

On the other hand, early 5.Xbeta versions of unRAID did not recognize the proper pre-clear signature on a 3TB drive. (so they cleared it again)

 

Joe L.

So when I add this drive back to my array, will I not still have this problem?  I ask because the first time I added this drive, even though the array must not have recognized the preclear signature; it did clear the drive (took about 9 hrs) and add it to the array.

 

From these replies it sounds to me like I could well have this same problem - is it clear what the problem actually was?

 

Sorry if I am being obtuse, and thanks again. 

Well I used the dd command as suggested.  I stopped the array and added the drive.  I then started they array, and the damn thing is clearing the disk again.  I am really not happy about this, but I guess we will see what happens after it is cleared.

 

 

My gut feeling is that I will have exactly the same problem.  Can someone please look at my posted syslog and see if there is some specific issue?  I really need this data drive.

 

Another symptom:  In the mymain/unmenu list, this new drive is listed as "unknown 3 T" in the "type" column.  I have never seen this before and it really seems to me this is a problem.  The clearing should finish in about 9 hours.

 

If this is a waste of time to wait; how to I stop the server from clearing the drive?  I am very worried about data loss now.

OK I am having _exactly_ the same problem.  Here are the last lines of the syslog:

 

Is this a bug?  What is the problem with disk 9?  If you look at the end of the syslog, from about 6 pm tonight, you will see the system has just finished clearing the disk, disk 9, and has a problem mounting disk 9.  Is this a bug?

 

Also, I am having a very slow webgui.  At this point it looks like I am not going to be able to connect to the main screen.  There is no response when I try to navigate to the main webpage until I get "the connection was reset."  Now what should I do? 

syslog-2012-03-08.zip

Can anyone suggest a path forward?

Did you press the format button?

 

I was completely unable to log into the array unmenu/main webpage - it would sit until the browser returned a timeout error, as I posted earlier.

 

So no, I didnt press the button - I could not access the page!

 

The next day I was able to find my laptop still had that webpage in its cache, and when I refreshed the page, I was then able to press the format button, and it added to the array.  I only had 5 min to do this before work so I didnt try copying any files.  This is hte only thing that allowed me to avoid literally pulling the plug on the server.

 

However I am pretty sure I will not be able to, as I have seen identical behavior at every step of the way both times I have gone through this process.  Using the 1.13 version of preclear.sh did not change anything at all.  There must be a bug, or several, related to this.

One bug is not explaining what you did. A few posts earlier you posted that you used the dd command and now you just posted that you used the newest preclear script. I clearly posted a couple of times to use one or the other. So, could you explain exactly what you did.

 

The server becoming unresponsive is pointing to bigger problems than just the drive not being accessed. I'm not sure what, but something is wrong to cause that.

Hi Lionel,

 

As I explained, I had the unraid main webpage cached in a different computer's webbrowser, and was then able to press the format button.

 

Formatting worked fine.  The array came up.  I copied a couple of files, and that did not work - again.

 

So, I rebooted the machine and then the new drive was recognized.  Thus, the handling of a 3tb drive is a bug.  I have the syslog attached and unfortunately I am too ignorant to figure out exactly what the problem is.  Maybe the inability to access the unraid main page is also a bug - the same or different, I do not know. 

 

Prior to this, originally I had followed the normal way to add a drive.  I precleared the 3 tb drive.  The report was fine.  I added it to my array, and the array did not recognize it, so it proceeded to clear the drive for 8 hours.  I then formatted the drive.  When the array came up, I copied some files, and unraid would refuse to send them to the new drive.  Then I went through the process again, except that instead of preclearing the drive again, I used the dd command.  This did not work; when I added the drive to the array it proceeded to clear the drive.  At that point I was really annoyed so I closed the web page to wait for the thing to finish.  I went back the next day and the web interface was very unresponsive; until I discovered I had the cached unraid main webpage available.  As I said, after formatting the array refuses to copy files to my new 3tb drive, exactly as before.

 

rebooting fixes the problem; but this is not part of the normal procedure for adding a drive to the array; thus a bug.