HI all,

 

I am running version 4.5.1, with 13 data discs and 1 parity drive.  I was copying data from disc 12 to disc 2, and halfway through, the copy stopped responding.  Disc 2 turned red, all others were green,  There were write errors on disc 2 AND disc 3, and both temperatures showed zero.  The web interface then hung up.  I was forced to do a hard reboot.  The array came up again, with the array started, and Disc 2 RED (all others green).  I rebooted again (from the web interface), and now all discs are BLUE.  I have tried a few more reboots, and everything is still blue.

 

Any advice on how to proceed safely?

 

Thanks!

 

 

Attach a syslog.

Syslog attached:

syslog.txt

Power down and run checkdisk on the flash in a Mac or PC.

Thanks.  I ran CHKDSK on the flash.  Some errors were reported, so I reran with CHKDSK /F.  A boot of UNRAID still shows all discs as blue.

 

New syslog attached.

syslog.txt

Un-assign the parity drive and start the array If the disks look normal and the contents are available then stop the array and assign parity. DO NOT format any disk.

Okay, it's fairly clear you've lost the contents of disk #2 and need to start over.    But if you're going to do that, I certainly wouldn't stay on a release as old as 4.5.  So do the following:

 

(a)  Note the serial number of the parity disk -- be CERTAIN you have this correct.

 

(b)  Copy your Pro.key file from the USB flash drive's Config folder to a safe place on your PC.

 

©  Reformat your USB flash drive in a PC with the correct label ("UNRAID");  then copy either the current stable release (v4.7) distribution to it or (what I'd suggest) the current v5 Release Candidate (v5 RC16c) to it.    RC16c is preferable, as (1) you'll be on the latest release; and (b) v5 supports larger disks in case you ever want to use a drive > 2TB.    When done, copy your Pro.key file to the Config folder.

 

(d)  Boot to the new USB drive, and assign your 13 data drives.  Then Start the array.    This SHOULD result in a good array.    Do NOT allow it to format any drives if it shows "Unformatted";  and do NOT do anything else if you see anything except a good start.  If anything isn't right, capture a picture of the Web GUI and post it here.

 

(e)  If the array started okay, confirm you can access it via Windows Explorer; then Stop the array and assign your parity drive.    Then Start the array and wait until it does a parity sync (MANY hours).    When that completes, do a parity check to confirm all went well (many more hours).

 

Done    ==>  AND upgraded to the latest release and ready for v5.0 "final"

 

 

NOTE:  There IS one thing you can try BEFORE the above.    Re-seat (unplug/re-plug) ALL of your drives in their slots (if you're using hot-swap cages) or all of your cables if they are directly connected.    Then try one more re-boot. 

Thanks.  But before I upgrade, I would like to get system back to "stable."  I am not convinced I lost data on Drive 2.  After un-assigning the parity drive, and starting the array, all of the data on the drives (including drive 2) looks OK (through windows explorer).  All data drives are green.  Once I  stopped the array and re-assigned the parity drive, now all are green EXCEPT the parity drive, which is BLUE.

 

My plan, perhaps not the best, is to get this array back to stable (even if on old UNRAID version), then build a new box.  Mine has been unstable since forever (mainly hardware/cabling issues).  Once I have the new array built, I will be transferring all data over.  I need to retire the multiple IDE drives.

 

So with all drives green, and the parity blue, I can just "Start will bring the array on-line and start Parity-Sync- I am sure I want to do this (checked)"

 

(I am used to the 15 hour parity-syncs )

 

Thanks!

 

... All data drives are green.

 

Ahh ... that's a significant change since your last status, where you indicated "... A boot of UNRAID still shows all discs as blue."      And since "... all of the data on the drives (including drive 2) looks OK (through windows explorer). ..."  it seems you can now read all of your data, so indeed you probably haven't lost anything.

 

 

My plan, perhaps not the best, is to get this array back to stable (even if on old UNRAID version), then build a new box.  Mine has been unstable since forever (mainly hardware/cabling issues).  Once I have the new array built, I will be transferring all data over.  I need to retire the multiple IDE drives.

 

It sounds like it's already "stable" ... except it doesn't have valid parity drive (it thinks it's a new drive).  Note that a parity sync will completely read every disk to compute parity.    Perhaps a better use of all these reads would be to backup all your data to new drives.    If you don't have backups of your data, you may want to get a couple large external drives 3-4TB units and remedy that NOW.  You could also use bare drives with an eSATA or USB docking station [e.g. http://www.newegg.com/Product/Product.aspx?Item=N82E16817153112 ]    (That's what I use for backups ... I store the drives in these:  http://www.newegg.com/Product/Product.aspx?Item=N82E16817997031 ]

 

THEN ... after you're well backed up ... you can go ahead and do a parity sync to provide fault-tolerance until you can get your new system built.

 

 

So with all drives green, and the parity blue, I can just "Start will bring the array on-line and start Parity-Sync- I am sure I want to do this (checked)"

 

Yes, you can do that ... but as I noted above, you may not want to 

Crashed during parity check- web not responsive

 

Crashed during parity check- web not responsive

 

As I noted above, generating and/or checking parity is going to stress all of the drives throughout their entire surface.    Given that you can read them in their current state, I would NOT do anything EXCEPT read the data and copy it to a backup (unless, of course, you already have backups of all the data).

 

If you don't want to buy a couple large external drives for backups; then shut down the server; and build your new one NOW.    Then copy your data to the new server one-disk-at-a-time.  [but you should still backup your data if it's important to you.]

 

Thanks.  Unfortunately I don't have 9TB laying around to back my data up - that is why I invested in an unraid server

I may have to rebuild now.

 

The parity rebuild keeps crashing.  And upon boot, the unraid server gives me the following:

"Pri Master Hard Disk: S.M.A.R.T..  Status BAD.  Backup and Replace.

Press F1 to resume.

 

Any ideas?

... Unfortunately I don't have 9TB laying around to back my data up - that is why I invested in an unraid server

 

Neither UnRAID nor any other RAID system is a backup.  Fault-tolerance does NOT replace the need to backup your data.    However, since you plan to build a new server to copy all your data to, you can re-purpose all the drives in your current server as backups      Just store them in these:  http://www.newegg.com/Product/Product.aspx?Item=N82E16817997031

 

 

I may have to rebuild now.

 

Definitely a good plan.

 

 

... Pri Master Hard Disk: S.M.A.R.T..  Status BAD.  Backup and Replace.

 

ASSUMING that the bad drive is your parity drive (I suspect this message is from your system's BIOS ... NOT from UnRAID)  then you could replace the parity drive.    But building your new server is a FAR better idea at this point.    Are all the drives still showing as green in the Web GUI (except, of course, the parity drive)??

 

Thanks.  Unfortunately I don't have 9TB laying around to back my data up - that is why I invested in an unraid server

I may have to rebuild now.

 

The parity rebuild keeps crashing.  And upon boot, the unraid server gives me the following:

"Pri Master Hard Disk: S.M.A.R.T..  Status BAD.  Backup and Replace.

Press F1 to resume.

 

Any ideas?

 

Are you getting into unRaid boot up when you get that or machine boot up?

 

Sounds to me like you are not even getting to unRaid startup. Your BIOS may be trying to boot from one of the Disks instead of the flash drive.

 

Go into your BIOS config and ensure your Flash Drive is the boot drive.

You may also want to take your Flash Drive to a Windows machine and run Checkdisk on it

 

It is booting from unraid.  The parity rebuild starts automatically.  Then a few minutes in I lose connectivity to web interface, then I cannot even ping.

 

Chkdsk shows no problems.

 

Trying to rebuild parity- at the rate indicate will be 275 hours.  Something is not correct.

 

While I love this product, my array has been unstable since day 1.

 

Anyone recommend a good motherboard, memory and power supply?  I am really over the instability of my current build.

 

 

 

  Sounds to me like you may have multiple issues with your current unRAID.  I would suggest looking at the SMART drive information in detail with no data transferring to, from, or within the array between drives.

 

  First, after booting unRAID, go to the web interface and CANCEL the parity check.  If the system is not stable or becomes non-responsive after this in the following steps I would look at your power supply, then your RAM...

 

  You will likely need to look at the SMART data for EACH drive, but based on your logs and descriptions,

  I would be most concerned with your;

    DATA disk2 - ST3500641A 3PM17RN6 (motherboard Primary Master IDE) - reported SMART FAILURE and redballed...

    DATA disk3 - ST3500641A 3PM103R2 (motherboard Primary SLAVE IDE) - on same bus, may have been a cause, or affected

  Also look closely at the Parity drive.  Look for errors such as bus errors, re-allocated sectors counts, re-allocated event counts, pending reallocations, etc.

 

  Assuming you are running unMENU...  In unMENU, select MyMain.

    Within MyMain, select Smart for the current view.

 

      This will give an easy place to see your SMART status for each drive.

 

  If you are running an older version of unMENU, without the MyMain plugin loaded...

 

      In unMENU, select Disk Management.

        Now you can select each drive, one at a time, from the drop down menu.

            Select Smart Status Report - The SMART status will be shown for the selected drive below.

 

              Look for errors in the SMART Error Log, such as bus errors, etc.  Pay close attention to

              Reallocated_Sector_Ct, Reallocated_Event_Count, and Current_Pending_Sector values

              (names will vary by manufacturer and model...)

 

            Select Short Smart Test to initiate a quick test of the selected drive.  WAIT ABOUT 3 MINUTES or longer if it says

              something longer than... Please wait 2 minutes for test to complete.

              3 minutes should normally allow a short test to complete on a good drive...

            Select Smart Status Report again - The SMART status will be shown for the selected drive below.

              Near the bottom you will see the results of the short test just completed under SMART Self-test log.

              look at the LifeTime(hours) column to see the time the test was run, the current one should match with the

              Power_On_Hours parameter.

 

            You can also select Long Smart Test for a much better and more revealing test if needed, it will take a long time...

             

 

Thanks for the response- I am not running unMenu.

 

The server did start to rebuild parity, then about about 10% through started getting Disc 2 errors.  So I believe I now have unstable parity, and a failing(or failed) Disc 2.  I am going to build a new server, with modern discs, and then transfer data disc-by-disc.  That being said, I have a feeling I may lose data on Disc 2.

Post SMART reports. What PSU model?

  Yes, I already believed you have a failing or unusably failed disk 2.  Your parity drive may be good, since I did not see any real concerns that might indicate a failing drive there, but the data is likely not correct now with all that has occured.

 

  Your disk 3 that is on the same IDE cable as disk 2 may have had buss errors, that may have either added to problems with disk 2, or been caused by the errors on disk 2.

 

  Since you do not have unMenu installed, you could add it for an easy way to see the SMART reports, and information on your drives.  The array does NOT need to be started or running to add unMenu, or utilize its functions.

 

  Alternatively, you could move drives (2 and 3 for testing, possibly also parity) to a Windows machine, and DO NOT allow windows to do any automatic partitioning, preparing, or formatting while adding the discs to a Windows machine.  You can also install CrystalDiscInfo in Windows,( http://crystalmark.info/download/index-e.html  ).  This is an easy to use, quick to see results program to look at SMART data on drives.  Since a drive does not need a reckognized file system, or even need a file system to work properly with the program, it is a good program to check the SMART data on drives.  (I run it on all my Windows machines...)

 

  You may also want to consider using unMenu regularly, and look at the smart data on your drives monthly... via the myMain - Smart tab.  That way you will know before you actually loose data, or have a system become unstable due to drive failures, that you need to replace a failing drive.

 

  The new simple features plug-in may also have SMART capabilities, but since I do not use it, I am unsure and unaware of what it actually offers.  But there are some users that run unMenu even though they are also running simple features.

 

  I really think it may be likely, that disk 2 may have been the main, if not only cause of your system being unstable since you have noticed the problems.  But I have also seen some similar symptoms actually casued by the 'other' drive on an IDE cable, so I would not trust your disk 3 yet until looking at it more closely on the SMART data also.  You may also want to actually verify the jumper settings on both disc 2 and disc 3  Just to make sure they are properly set, and it may also help to make sure the drives are actually set for MASTER and SLAVE instead of using the cable select option on both drives.  I have seen on some combinations of drive controllers and drives, that the cable select option may in rare situations cause problems with drive connectivity.

 

  Just trying to help you not loose any more data, than you may have already lost.