MD1000/PERC5e, trying to understand my predecessor

Servers

Servers
Information and ideas on Dell PowerEdge rack, tower and blade server solutions.

MD1000/PERC5e, trying to understand my predecessor

This question is not answered

Hello all,

I am coming onto a workplace with hardware arlready installed and configured, but something is not quite right with it.

We have an MD1000 connected to dell 2950 with the perc5e.  The md1000 has 12-500gig drives

8 disk is configured as raid 50 with 2791.50GB total disk space.  This was then "mounted" I'm not sure if I use the right term there, on a windows 2003 storage server as drive E.  Unfortunately, due to limitations of windows 2003, only 2 terrabytes was used leaving about 740gig unallocated and unusable.  From what I've read, I won't be able to reconfigure the raid without rebuidling it right?  What is the best possible solution to take so that I can utilize the 740gig.  Rebuilding seems out of the question since I don't know how long it would take to do it and even with backup, I don't feel confident enough with just a backup. This storage is used by programs reading and writing to it all day and night and the backup may not back everything up while the file is in use.  It's hard to bring it down much less down for more than a day.

Is the 740gig lost for good or I can do something about it?

 

All Replies
  • I believe Server 2003 SP1 takes that limitation away. Check out this link to see if there is anything you can do

    http://www.microsoft.com/whdc/device/storage/LUN_SP1.mspx

     

    In terms of the raid there is really nothing you can do without rebuilding it. the only thing I could think of is if you were to purchase Disk manager from Paragon Software to split up that large partition into two, but I would call Paragon and find out to see if this would work for you.

  • Windows 2003 SP1 indeed can use disks over 2TB, but there is a catch:

    To be able to use the space after the 2TB point on the disk, the partition table needs to be GPT. Windows (and Linux) can only convert to GPT if the disk is completely empty (no partitions on it).

     

    So, assuming your server is running SP1 or later, to use the remaining 700GB you will need to first check if the disk isn't already GPT. To do this, in disk management, right click the disk (not the partition), and select properties. Look under the volumes tab. It should report the partition style. There are 2 options; MBR (Master Boot Record) or GPT (GUID Partition Table).

    If the disk is currently on MBR the steps will be:

    - back up all the data on your E-drive

    - go to disk management and delete the E-partition

    - right click the disk (probably "Disk 1"), not the partition, and select "Convert to GPT"

    - recreate the E-partition (may as well go max size) and format it

    - restore your data

    Member since 2003

  • - right click the disk (probably "Disk 1"), not the partition, and select "Convert to GPT"

    Ah...that's the trick.  I've read both replies above and it kept on creating mbr partition and chopping off at 2tb, but righclicking on the drive, not the partition and say convert to gpt did the trick.  It recombines it.  Anyway, we have a second  md1000 that wasn't previously utilized before, and I was able to format it as 1 large chunk.  So I'm going to set up the right way, copy over the data from the first md1000 to the second, and redo the first one to recover some space.  The trick now is to copy everything over making sure that all the rights and security remains the same, then I'll swap the drive letter and re-establish all the shares so it's transparent to the users and programs using the data.  I've done this before using robocopy.  If you have any better ideas let me know.  Oh and probably do a backup first before I do all this.