What is the easiest way to upgrade an existing mirror to new 2 bigger drives?

What is the easiest way to upgrade an existing mirror to 2 new bigger drives?  I currently have a production server that is a Dell Poweredge 840 with a SAS5/IR Controller card that has 4 Sata ports with only 2 being used by 2 250GB sata drives.  I run Zimbra Server on this unit and the 250z are filling up quickly so I purchased 2 750GB sata drives from Dell and now I am stumped on how to smoothly and easily transfer the existing mirror to these 2 new drives.   Dell is of know help with this even at Enterprise level (Silver).  There has got to be a way to do this given that anybody who has a mirror config server would have to upgrade at some point to bigger drives?  Any help from the Experts would be totally 100% graciously appreciated.
LVL 1
rwjblueAsked:
Who is Participating?
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

uucknaaaCommented:
Hi
I've been reading through what information I can find on this adapter on Dell.  You are right, there is much.  I did find the manual.  Been reading and haven't found anything specific yet.
Here's what I'm thinking.  Could you replace on of the drives in your existing mirror with one of the new ones.  Even though it's bigger, my experience is most system will let you do that and it just sizes according to the smaller drive.  Then synchronize the mirror and replace the smaller element with the other big one.  Then upgrade the size of the mirror.  That sounds like a lot and many ifs.
The other option would be to create a new mirror on the 2 ports not being used.  Then transfer the data over to the new larger mirror and put the new virtual drive in production.
Are these things you've been thinking of?  I'll keep reading while waiting your response.
 
0
rwjblueAuthor Commented:
Hey Uucknaaa, Thanks for you reply.  I like the last option you listed:

"The other option would be to create a new mirror on the 2 ports not being used.  Then transfer the data over to the new larger mirror and put the new virtual drive in production"

Now with this solution are you saying keep the existing mirror, shutdown the server, place the new 750z in the other available 2 sata ports, boot up, go into the controller config, create another mirror for the new 750z, and then I guess use 3rd party software to transfer data from the old mirror to the new one?   Thanks again for your help any suggestions is greatly appreciated.


0
MCode151Commented:
Snips to support uucknaaa's 2 ideas:

Note the biggest issue/reason for backup should something go wrong:
Unlike most controllers [...] you cannot start with a single drive and create a RAID 1 array without deleting data on both disks.

Also complaints that the automatic rebuild doesn't happen, have to start it manually.

http://support.dell.com/support/edocs/storage/RAID/SAS5iR/en/UG/HTML/index.htm

Integrated Mirroring (IM) provides protection against the failure of a single physical disk. When a disk fails, the physical disk can be replaced and the data re-mirrored to the physical disk, maintaining data integrity.

In the event of a physical disk failure in an IM virtual disk, you will need to replace the disk and resynchronize the virtual disk. Synchronization will occur automatically on replacing the physical disk using the following steps.

   1. Replace the failed physical disk with a blank disk of the same type and capacity. (I would assume greater ok as it's usually possible with most RAID 1 setups.)

   2. Check your management application or the BIOS Configuration Utility (Ctrl-C) to ensure synchronization started automatically.
   
NOTE: The Dell SAS 5/iR controller only supports migration of consistent virtual disks and does not support hot migration of virtual disks.
 
NOTICE: If the physical disks of a virtual disk are removed and the virtual disk's configuration is subsequently deleted from the SAS 5/iR controller, the physical disks show up only as simple disks with no RAID association if they are placed back onto the same SAS 5/iR controller. Once the virtual disk is removed from a SAS 5/iR controller using the BIOS Configuration Utility (regardless whether the physical disks members are present), the virtual disk cannot be restored.

Thus don't remove both physical disks/the virtual disk (unless following the migrate steps) because 'Creating a Integrated Mirroring Virtual Disk' contains:
Note: Data on both disks will be lost. Dell recommends backing up all data before performing these steps.

See review 2 & 3: http://reviews.dell.com/2341/341-4341/reviews.htm
(2 is also stating that the synchronization doesn't happen automatically)

Cons1.
If one out of your RAID 1 SATA disks fail and you replace it with another drive, the RAID 1 virtual drive is not automatically rebuilded. Sometimes you even cannot use OpenManage and force the 5/iR to rebuild it. All you can do then is to backup everything and proceed to step 2 (building a new RAID 1 virtual disk), see below.

2.
If you build a RAID 1 virtual disk out of a single disk, all your valuable data will be ERASED.

Cons
Unlike some third-party SATA RAID adapters, you cannot start with a single drive and create a RAID 1 (mirrored) array without blowing up the original drive. This means it's good only for original installs, not easy upgrades.

Solution A:
-If you can create a backup first do so.
-Pull 1 disk (Put in safe place, if you couldn't create a backup above might have to use this as a single drive somewhere to copy/restore data onto a 2x750 array should all else fail.)
-Stick in a 750GB
-If not noticing a lot of HD activity after sometime (automatically rebuilding) see bios (ctrl+c) for synchronization options.

Should you have to totally rebuild the mirroring configuration, it will wipe all data on both disks involved. (Which is why you have that backup/disk1 in a safe place.)

Solution B:
Or you can take the other approach all together and add the 2x750's as a 2nd virtual disk later removing the 2x250's once you copy data over. But how your booting becomes an issue/might do it from a live CD.
http://support.dell.com/support/edocs/storage/RAID/SAS5iR/en/UG/HTML/yd912c51.htm#wp1057659

Unfortunately it appears you can't create 1 virtual disk consisting of 2 x RAID1 you can only span virtual disks across 4 drives in RAID 0 mode, and only the PERC 5i has more modes.
0
Newly released Acronis True Image 2019

In announcing the release of the 15th Anniversary Edition of Acronis True Image 2019, the company revealed that its artificial intelligence-based anti-ransomware technology – stopped more than 200,000 ransomware attacks on 150,000 customers last year.

uucknaaaCommented:
Hi all ,
MCode151 sure left a great post.  There are many points there I learned from and issues were brought up that I hadn't thought of.  Thanks for your help!
I'm trying to think of a way to do this and explain it simply.  Here's the way I see this happening:
Create the new mirror and get everyting ready for the data transfer.  At that point, you could use something like Norton Ghost to make an exact copy of the disk you are migrating.  If you are useing this volume to boot from, the Ghost image will retain all of the boot options necessacy for your server.  If it's just a data drive, it may be as easy as doing a copy.
rwjblue, when you've had time to absorb all this, get back with any questions you might have and we'll get your server upgraded.
Waiting your response.
0
rwjblueAuthor Commented:
Uucknaaa and Mcode151 thanks so much for your responses.  I guess from I have digested is that perhaps the best avenue would be to take one drive out and replace with a 750gb drive and rebuild mirror, then after this process is finished replace the other drive with a 750gb and rebuild the mirror - Now my question in regards to this step does the controller bios options let you fill the drive up? so I am not capped at 250gb.  I was not sure if this was stated whether or not the controller options would let me perform this task or if its capable.  I have absolutey no issues with this method given that the drive well not be capped at 250GB.  Can somebody please clarify this.  I think we are getting close to a final solution here.  Thanks...........
0
MCode151Commented:
I'd rather add it as a 2nd volume and put just the Zimbra data on it, but I'll cover the other aspect/Q's first below.

The documentation is pretty useless in specifying any expand volume commands. I have to with my much newer RAID 5 & RAID 10 controllers. (Automatically expanding when all driives in the array have space isn't a common feature and a few devices which do so aren't typically really RAID, they're JBOD with redundant algorithms). Worse case I suppose you could do the process twice, get it back on the 250's, then switch to the solution B.

I'm getting the gut feeling that you'll end up having to to use the clone method. Wish we could be more help, hopefully someone who owns one will come along.

Again a OS level or Zimbra data backup might be prudent; even if it's just an rsync -a or tar copy of /opt/zimbra.

--- >>> The other possibility (what I'd probably do if I had extra drives/backplane) is just add the 2 750's as a store2 & index2, OR move your store and index to the 750's leaving the 250's for the OS and rest of the Zimbra software.

TO ADD:
Realize blobs are NOT moved in this process, you're only pointing at the location to store any NEW data.

Each Zimbra mailbox server is configured with one current index & message (NE can add secondary HSM volumes).

When a new message is delivered or created, the message is saved in the current message volume. Additional message volumes can be created, but only one is configured as the current volume where new messages are stored. When the volume is full, you can configure a new current message volume. The current message volume receives all new messages. New messages are never stored in the previous volume. A current volume cannot be deleted. If there are accounts with messages referencing a volume it should not be deleted until you point the volume path to wherever you've now moved the items.

Each mailbox is assigned to a permanent index directory on the current index volume. When an account is created, the current index volume is automatically defined for the account. You cannot change which index volume the account is assigned. As volumes become full, you can create a new current index volume for new accounts. When a new index volume is added as current, the older index volume is no longer assigned new accounts. Index volumes not marked current are still actively in use as the index volumes for accounts assigned to them. Any index volume that is referenced by a mailbox as it's index volume cannot be deleted.

The Network Edition has HSM (Hierarchical Storage Management) which involves moving data to other storage locations automatically after x amount of time. Messages and attachments are moved from a primary volume to the current secondary volume based on the age of the message; completely transparent to the user.
For instance: After 30 days messages move from your fast (and usually more expensive) SCSI disks, to a SATA array so that you can keep storage costs down.

The admin console > server > volumes tab is straight forward, make sure you have permission to write to the target location - see also: http://wiki.zimbra.com/index.php?title=CLI_zmvolume
(zmvolume -l and -dc arguments display your config)

or in the DB that's:
mysql> select * from zimbra.volume;


TO MOVE STORE:

Mount the 2nd volume in the /etc/fstab, stop zimbra, copy (with all flags/permissions) data to /mnt/zimbra/store, then create a symlink /opt/zimbra/store > /mnt/zimbra/store.


To MOVE ALL:

Or copy & mount  the entirety of /opt onto the 2nd set of drives, note that Zimbra doesn't like /opt/zimbra as a symlink http://bugzilla.zimbra.com/show_bug.cgi?id=14725 Fix is just a warning, so actually mount /opt separate, or use mount with the --bind switch to link two directories.

Now your Zimbra data & sofware is 'separate' from your OS. (Use the ./install.sh -s trick should you ever reinstall the OS completely or desire to switch platforms.)
0

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
uucknaaaCommented:
Hi

I agree that the safest way to accomplish the migration would be to create a new mirrored volume with the two 750 GB drives.  Then move the data from the old mirror to the new.  The only way I'd attempt the other solution, especially with production data, is if you couldn't have any down time and needed the data to be accessible all the time.

There has been a lot of great discussion here.  I'm going to save a link to this for my own reference.

Later ...
0
rwjblueAuthor Commented:
I appreciate the feedback once again.  It seems to me this is getting way to complicated.  I am new to linux and very familiar with Windows and recall my days of using Ghost and being able to clone a drive then using software to release the remaining size of the drive for use.   Why not clone the existing drives maybe with clonezilla or some other linux based cloner and then resize the drives.  I am just throwing this out there as this something I googled and found on the net.  There has got to be an easier way for this to be done.  I can't imagine Sys Admins having to perform this process for simply upgrading the drives on a mirror to a bigger size.  I am not in any way arguing but simply suprised that what seems to me to be a task that most admins run into frequently is so complicated and resulting in huge down time.  I think if this is ultimate answer then there is a possibliity for somebody to retire by inventing a way to make this easier.
0
rwjblueAuthor Commented:
I am giving both you guys the points for your help.  I appreciate it.
0
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
System Utilities

From novice to tech pro — start learning today.