20 Replies Latest reply on Sep 30, 2010 7:51 PM by JEShort01

    How do I remove one failing hard drive from raid set and replace with new one

    rowby Level 1

      Last Friday apparently one of my raid drives started failing.

       

      As I mentioned on this forum I started getting continous beeping.

       

      I was finally able to get the raid working at a degraded level.  I ordered a replacement hard drive which is arriving today.

       

      (In the meantime I made twice daily backups of my work.....)

       

      Below was the message I got from the browser based raid software:

       

       

      Blahblah 09    1000.2GB   RaidSet Member SamSung HD103SJ

      Blahblah 10    1000.2GB     Free                 SamSung HD103SJ

      Blahblah 11    1000.2GB   RaidSet Member SamSung HD103SJ

      Blahblah 12    1000.2GB   RaidSet Member SamSung HD103SJ

       

      (See this earlier thread if you wish!)

      http://forums.adobe.com/thread/727867?tstart=0

       


      At one point when I did the checked the browser interface I saw the message Failed and Degraded

       

      As I said, I was able to work over the weekend on the degraded system.

       

      This morning I got the beeping again and did the rescue and now I am running a "full raid" without the notice that one raid was "Free".

       

      In any case, the new hard drive is arriving today.

       

      What steps should I take to incorporate the new drive into the raid system.

       

      I have one OS drive

      and 4 tb raid drives.   One needs to be replaced with the new one that I am getting today.

       

      Thanks

       

      Rowby

        • 1. Re: How do I remove one failing hard drive from raid set and replace with new one
          ECBowen Most Valuable Participant

          You first have to isolate drive ID 10 so you can replace the correct drive. I would offline the raid then shut the system down. Trace to port 1 on the card since most cards are labeled ports 0 to 7. If you have a data cable attached to port 1 then pull that cable from the drive. Start up and see if drive ID 10 is no longer showing. If it's not showing then you know that is the drive. Shut down and replace the drive. Now if you have cabled into ports 4 though 7 then it should be port 5. Follow the same procedure to isolate.

           

          Eric

          ADK

          • 2. Re: How do I remove one failing hard drive from raid set and replace with new one
            rowby Level 1

            Thanks Eric!

             

            That makes sense.

             

            When the drive arrives I may have more questions.  But it seems like a logical approach!

             

            Rowby

            • 3. Re: How do I remove one failing hard drive from raid set and replace with new one
              Harm Millaard Level 7

              Rowby,

               

              Like Eric said, the most difficult part is isolating the failing disk 10. What can help is using your hands to determine which drive is cooler than the rest and does not vibrate lightly, like the other disks do. Once you have found that disk, disconnect the SATA cable and then restart. If the identification process went correctly, you will still see the same degraded info.

               

              Now turn off your PC, remove the failed disk and insert the new one, reconnect the cables and restart. It will still show a degraded array. Does not matter, since it is now easy to solve this. In the BIOS or the Raid Storage Manager (from Windows), define the new disk as a Hot Spare. Let it finish this process. Then use Expand Raid Set to include this Hot Spare into the Raid. It will rebuild automatically.

               

              That's all, but it may take some time. Good luck.

              • 4. Re: How do I remove one failing hard drive from raid set and replace with new one
                JEShort01 Level 4

                Rowby,

                 

                To definitively know you are changing the correct drive, I would determine the serial number of the drive that is flagged by the Areca controller as bad using Areca's browser tool. Once you know the bad channel ID, you can see what the drive serial number by:

                1) Open Raid Storage Manager

                2) Click on Information / RAID Set Hierarchy

                3) Find your drive serial number (match slot number) under the appropriate enclosure reports under the Model column

                 

                Apologies if this does not work for your Samsung drives; this does show the serial numbers for Western digital, Intel, and Seagate drives that I have used personally.

                 

                Jim

                • 5. Re: How do I remove one failing hard drive from raid set and replace with new one
                  rowby Level 1

                  Hi Jim,

                   

                  I will be doing the swap this weekend (in the meantime doing serious backups) since I don't have time to rebuild my raid until then. 

                   

                  I am happy there is some kind of browser to hopefully help me identify the various drives.  I would hope that, like the Western Digital, that the Samsung drives can be identified by their serial number.

                   

                  Stay tuned!


                  Rowby

                  • 6. Re: How do I remove one failing hard drive from raid set and replace with new one
                    Harm Millaard Level 7

                    The drawback of Jim's approach is that you need to remove all your disks, just to be able to read the serial number on each disk. Yes, the webinterface shows the serial numbers for Samsung disks, but to confirm it, you need to uninstall them. The manual approach makes more sense I think. I once had one disk fail out of 12 in the array. I did not even consider to uninstall all of them. The manual approach worked easily, even with 12 disks. With 4 it is even simpler.

                     

                    For the future, mark all SATA cable with a number that reflects each port. That makes identification quite easy. Then if the disk on port 6 fails, all you need to do is find the cable marked 6.

                    • 7. Re: How do I remove one failing hard drive from raid set and replace with new one
                      Powered by Design Level 4

                      Great idea Harm.

                       

                      Little things like that can save you so much time if needed.

                       

                       

                      GLenn

                       

                       

                      Harm Millaard wrote:

                       

                      For the future, mark all SATA cable with a number that reflects each port. That makes identification quite easy. Then if the disk on port 6 fails, all you need to do is find the cable marked 6.

                      • 8. Re: How do I remove one failing hard drive from raid set and replace with new one
                        JEShort01 Level 4

                        It seems that my suggestion regarding checking the drive serial number of a drive to be replaced has been misinterpreted. You do not need to remove all drives and read the serial number on each drive, but rather need to read the number on the drive that has been flagged as having issues. So, the only potential drawback would be the extra 5 minutes or so needed to use Areca's utility and verify a single serial number, and in my opinion this is good insurance!

                         

                        Jim

                         

                        • 9. Re: How do I remove one failing hard drive from raid set and replace with new one
                          rowby Level 1

                          Thankt for that clarification.

                           

                          One additional question.


                          After I restore my raid, I want to reinstall windows 7 completly.

                           

                          I will first of course do a backup of my raid onto separate backup drive(s).


                          But when after I reinstall Windows7  -- I assume I will be thenable to restore the same raid and all of its contents once again -- as long as I use the same settings as I did with my current raid.

                          • 10. Re: How do I remove one failing hard drive from raid set and replace with new one
                            rowby Level 1

                            I just now went into my archttp browser based areca raid interface and looked at System Events information.  Unfortunately the adobe form will not allow a spreadsheet so here are some screen captures.

                             

                            You can see in the "raid fails image" the events that led to the raid failing about a week ago.


                            I was able to rescue the raid and in fact all 4 drives are "working" again in the raid, and have been working for the past several days.   See the Raid set hierarchy configuration capture....

                             

                            See today's log (which is similar to previous day logs) with lots of Time Out Errors.

                             

                            Of course those time out errors as well as the hard drive "failure" concern me, which is why I am going to replace the bad drive sooner rather than later.  I plan to start today if possible, rather than waiting for the weekend.

                             

                             

                            I am reinstalling windows 7 for a different reason.  (I have been having a problem getting a 3rd party after effects plugin to work, even with the help of the plug-in's tech support.  I've reinstalled the complete Master Collection and the plug in continues to crash After effects -- even though it is CS5 complatible and the plug-in tech support has tested it in their system with no problems and have had no problems from their many other customers.  So I figured I might as well bite the bullet and do a windows reinstall too, as long as I am going through this process.)

                             

                            So my game plan is to:

                             

                            1)  Backup my raid.

                             

                            2)  Replace the bad drive and restore a healthy raid.

                             

                            3)  Reinstall windows.(Will I lose the raid contents when I reinstall windows?   I bet there are tricks that I can use to simply restore the raid with all of its contents and not rely completly on the backup.


                            Rowby

                             

                             

                            P.S.

                             

                            When the raid failed below is what the configuration showd (My blah blahs)

                             

                            Below was the message I got from the browser based raid software:

                             

                             

                            Blahblah 09    1000.2GB   RaidSet Member SamSung HD103SJ

                            Blahblah 10    1000.2GB     Free                 SamSung HD103SJ

                            Blahblah 11    1000.2GB   RaidSet Member SamSung HD103SJ

                            Blahblah 12    1000.2GB   RaidSet Member SamSung HD103SJ

                             

                            P.S.P.S.

                             

                            Sometimes I think I spend more time creating screen captures for these Adobe forums than actually editing my videos! 

                            • 11. Re: How do I remove one failing hard drive from raid set and replace with new one
                              Harm Millaard Level 7

                              Please tell me how to read the serial number from an individual drive rather easily, without un-installing them:

                               

                              P1000162.jpg

                              • 12. Re: How do I remove one failing hard drive from raid set and replace with new one
                                rowby Level 1

                                >>>

                                For the future, mark all SATA cable with a number that reflects each  port. That makes identification quite easy. Then if the disk on port 6  fails, all you need to do is find the cable marked 6.


                                <<<

                                 

                                Great Tip Harm -- one you should add to one of your Raid articles.

                                 

                                >>>>

                                 

                                Rowby

                                • 13. Re: How do I remove one failing hard drive from raid set and replace with new one
                                  rowby Level 1

                                  Attached is a photo of my current raid configuration.   As it turned out the cables had numbers already on them, and I plugged them in the same order into the raid hard drives (from the ARECA controller).

                                   

                                  P1 ---->  Drive 1

                                  P2 ---->  Drive 2

                                  P3 ---> Drive 3

                                  P4 ----> Drive 4

                                   

                                  See attached photo.

                                   

                                  So I would say it's reasonably "safe' to assume that Drive 2 is the one that is failing or is ready to fail.

                                  • 14. Re: How do I remove one failing hard drive from raid set and replace with new one
                                    rowby Level 1

                                    Reading this thread, I see I forgot to include "Today's log" in one of my earlier posts".

                                     

                                    Please see attached.

                                     

                                    Lots of "time out errors" but perhaps that is "normal" at startup????

                                     

                                     

                                    IN the meantime I am copying all of my files from the raid onto my external backup....  It's going to take several hours!

                                    • 15. Re: How do I remove one failing hard drive from raid set and replace with new one
                                      Harm Millaard Level 7

                                      You can never be too safe with data, but if you follow the hot-spare, extend raid set and rebuild sequence, it is not necessary. Your data will be reconsructed automatically. That is the reason you use a raid3/5.

                                       

                                      Your log file is a definite show of a serious disk problem. Replace it, create a hot-spare, extend your raid set and rebuild (it should be done automatically) and you are ready to roll.

                                      1 person found this helpful
                                      • 16. Re: How do I remove one failing hard drive from raid set and replace with new one
                                        rowby Level 1

                                        Thanks Harm,

                                         

                                        I am still doing the total backup.

                                         

                                        Should I rebuild the raid first

                                         

                                        OR

                                         

                                        Should I reinstall windows 7 first. and then do the raid fixer-upper.

                                         

                                        Rowby

                                         

                                        P.S.  Can I add the "hot spare" on Saturday. I just ordered another drive and it won't get here until Friday.  I assume I can.

                                        • 17. Re: How do I remove one failing hard drive from raid set and replace with new one
                                          rowby Level 1

                                          I remember when I created the original 4 TB raid, it took almost 24 hours to "build it"

                                           

                                          Are there ARECA configuration files on my OS "C" drive that I should copy so that I can jump start the rebuilding of the raid.

                                           


                                          SInce all of the data is currently on the 4 hard drives -- it would be nice to just "repoint" to them wiht some simple existing config file and all will be restored.

                                           

                                          I realize there's the replacement of the one failing drive that I need to deal with at some point, but any discussions about saving the almost 24 hours of a complete rebuild would be nices.

                                           

                                          Yes, I am still backing up my data files to my external backup drive.


                                          Thanks

                                           

                                          Rowby

                                          • 18. Re: How do I remove one failing hard drive from raid set and replace with new one
                                            Harm Millaard Level 7

                                            Rowby,

                                             

                                            You currently have IIRC your background priority set to 20%, so it will take some time to rebuild, but likely much less than 24 hours as for a new build. Even with 20% set, I guess it will be done in less than 12 hours, but you can set the background priority to 50% and it will be done much faster at the detriment of access to your raid. Background rebuilding is intended to give you access to all your data files and continue work, while the data is being rebuild. You can continue your work, so whether it takes 8 or 12 hours to regain full intergrity of your data is not all that relevant. Your will have slower response than with a full functional raid, but still has more speed than a single disk.

                                            • 19. Re: How do I remove one failing hard drive from raid set and replace with new one
                                              rowby Level 1

                                              Harm,

                                               

                                              I am away from my computer while it is making a backup.

                                               

                                              I will be back in front of it in a few hours.

                                               

                                               

                                              Maybe you already have an article on this, but where does my data actually reside.

                                               

                                              Does it start on "Drive 1 and then at some point starts to go to "Drive 1"

                                               

                                              I guess what I am saying, when I replace my second failing drive, and assuming there is data on it, is there duplicate data on other drives -- and that's how it's all re-crecreated?

                                               

                                              Again, I bet you have an article here that explains all this.

                                               

                                              Rowby

                                              • 20. Re: How do I remove one failing hard drive from raid set and replace with new one
                                                JEShort01 Level 4

                                                Harm,

                                                 

                                                Regarding your comment:

                                                 

                                                ***

                                                 

                                                Re: How do I remove one failing hard drive from raid set and replace with new one

                                                Please tell me how to read the serial number from an individual drive rather easily, without un-installing them:

                                                 

                                                ***

                                                 

                                                If you select the proper drive to change out, you only need to remove one drive and look at its serial number...

                                                 

                                                Step 1: Identify bad drive serial number using Areca's tools

                                                Step 2: Turn off the computer

                                                Step 3: Remove what you think is the bad drive based on following your numbered cable method, marked hot-swap bays, whatever

                                                Step 4: Verify that the serial number matches the "bad drive" serial number from step 1; if it does great, proceed; if it does not match, go back to step 3

                                                Step 5: Change out the CORRECT drive - that's the bottom line for this whole procedure

                                                 

                                                Cheers,

                                                 

                                                Jim