Poweredge 2950 Replacement Drive

Hi Again

I'm also struggling to understand how the RAID config  fits into this disk configuarion as I didnt set it up.

Along with the image above this is also a screen dump of the RAID config

This is the disk view in windows. The G drive isnt part of the internal setup. Its an external USB.

Can anyone possibly give me some pointers on how the disks shown in My Computer are fitting inot the RAID config?

Its driving me mad.

You have “at least” 5 drivers.   A RAID 1 is always build with 2 HDDs and a RAID 5 with 3 or more but, I think yours have 4 because of the standard HDD sizes.

Judging by the sizes: I think disks C & E are partitions on your RAID 1 volume and disk D is your RAID 5.  You can check this out on Windows “Computer Management” / “Disk Management”.

Disk E is labeled “Spare” but this does not mean it is a hot spare, the partition name is Spare, that’s all.   You do not have a hot spare.

I am not sure what you want to do: “Can someone advise on how I will go about removing the current C: OS drive and putting the new one in?” . . . can you clarify?

You cannot convert a RAID 5 into a RAID 1 because that operation will reduce the capacity of the volume.  You currently have a 836GB 4 drives RAID 5 and the RADI 1 volume you can build with those drives (2 of them) will be about 280GBs. 

If your data is backed up you can delete the RAID 5 volume, create a new RAID 1 and copy the data back.

You are correct in your assumption of our setup. I have since found out how to find out what was assigned where.

I've actually decided that i'd rather switch my RAID around than buy replacements.

I want to leave the C: and E: as they are on RAID 1 but I'd like to change the RAID 5 to RAID 1 and free up two disks of which one I would like to use as a hot spare and have the other one as a spare for if any disk fails in future.

Is it difficult to change the RAID 5 to RAID 1. I'm not even sure the D: DATA needs to be RAID as its backed up each day and there is only 20gb of data to restore.

can you advise please?

Thanksfor your help

Once I've ensured my data is backed up and deleted the RAID 5 do I create the RAID 1 in Dell Openmanage?

Once the RAID 1 is set up Will I then be able to create one of the spare drives (I will only be using 2 of the 4 drives for the RAID 1) as a global hot spare?

Do I need to do anything with the surplus drive?

If I decide to just have normal drives (instead of RAID 1) for my Data drive do I need to attach the drives to the SCSI (Non RAID) Card rather than the PERC 6i Controller (RAID) or will this happen automatically?

Yes, you can create the RAID 1 volume using OpenMange.

Once the RAID volumes is created you will be able to create a partition from the OS as the ‘spare drive’.

You do not need to do anything with the surplus drive but, if you do not have any plans for that drive you may think about setting it up as a second hot spare . . . you never know what may happen !!

With the PERC controller you will not be able to see the drives if they are not configured as RAID volumes.  An alternative to this limitation is to RAID 0 on each drive.

It seem you have enough hardware to create RAID volumes and to have hot-spares.  I would recommend you take advantage of that, it will save you time (money?) being able to keep operating even if a drive failed . . .

Thanks again

If I do not configure the drives as RAID Volumes will they show in Windows? If not how can I get them to show in windows?  When you say to RAID 0 each drive will this not take up all of my spares?

My main concern  (and is why I am doing all this work)  is if there was a double operating system array disk crash or corruption, meaning I have to rebuild the server from scratch (I dont have the time). I have an Acronis image of the virtual disk that holds the OS partition that I would hope to restore in those circumstances but not sure how I would go about doing this. Would I first need to make sure the new drives were RAID before restoring the image? (the version of acronis does support RAID)

I'm just trying to cover every possibility so that in the event of a crash I replace the drives, restore my acronis image, reboot and off we go again..

I never worked with Acronis.
The PERC controller only shows RAID volumes to the OS, if a disk is not configured the OS will NOT see it.
A RAID 1 volume already offers protection against hardware failure.  RAID 1 = mirror = all data copy twice (one is each disk).  If a drive fails the system will keep working with only one drive (no impact to your operation). If you have a hot-spare, the system will automatically rebuild the RAID 1 using it, so you are back on step 1.  On the mean time, you can replace the failed drive with a new one and the system will rebuild back to this new drive to keep the RAID volume on the same physical location.  Unless the OS gets corrupted for something not-hardware related, Acronis should not even get in the picture.
If you are a beginner on RAID this article may help you understand it better http://en.wikipedia.org/wiki/Redundant_array_of_independent_disks

Thanks for your replies both of you.

Cloning sounds a good idea. Its likely I'm going to have an OS spare to facilitate doing this.

something I have considered though is this. If its Windows corruption thats the problem I'm led to believe this will replicate to the other disk in the RAID meaning both disks will contain the faulty windows installation. If I have a clone then I'm ok to repair it but which disk would have the priority because I wouldnt want to install the clone only for it to be overwritten by the faulty install.

Do I need to run chkdsk before creating the Acronis image?

What do you mean by tombstone life?

Can I pull the disk without shutting down the server?

If you have raid 1, the odds of two disks failing within say two weeks is astronomically low, massive power surges, lightning, fire or theft is far more likely. Corruption/virus/user caused issues would be another matter, much more likely to happen.

At a client this past weekend I created clones of there Active Directory servers... With raid 1, you could do it  with Acronis, but the simple way is to order a spare disk for each server requiring an image, pull one of the members disks of the raid 1, replace it with a spare, which will rebuild. The pulled disk is an exact clone for use in an emergency restoration ( AD does have a tombstone life which needs to be considered, as does the number of AD changes between "cloning") . I work mostly with small businesses, so a cloned disk every couple of months, along with tape backups is all which is needed for emergency restores. If more frequent cloning is needed, the low MTBF of the backplane and drive connectors reinsertion needs to be considered.

I do not trust images, so with Acronis I generally hang a cheap SATA drive on the Sata interface, clone the array to it (after pulling the network cable to stop replication, if AD), reboot to verify the SATA boots to console, shutdown, disable or remove the SATA drive.  If you trust image files, at least run chkdsk on the server to be imaged.

" I'm led to believe this will replicate to the other disk in the RAID meaning both disks will contain the faulty windows installation." ...Yep, if a file is corrupt, you have a corrupt file on the both disks. If a physical error, on the first disk written to causes the corruption, it should not transfer to the second disk; physical errors on the second disk should also not cause a data error. The raid adapter checks for physical errors before committing writes.

"If I have a clone then I'm OK to repair it but which disk would have the priority because I wouldn't want to install the clone only for it to be overwritten by the faulty install." Very few times I had to restore a system using a  clone , so my procedure my not be the shortest/easiest, been a long time...Considering both disks have failed, or you have been hit with corruption or virus. Pull both disks, replace with clone, start up to make sure it boots ( will balk with the degraded message), place a  "good" disk back in ( which could be one of the corrupted pair, or a new replacement disk, if both disks failed), rebuild should start automatically but  newer adapters may try to protect the data on a good but previously used disk, and you have to intervene and force it to rebuild, by selecting the "good" disk, selecting rebuild. Had this happen over the last weekend while doing my cloning, got me a bit nervous, as the last time I had to institute a mirror from a cloned disk, after placing the second drive in the rebuild was automatic ( likely this  resulted from adding the second  disk while the server was shutdown) . I suppose the safest way would be to clear any configuration from a previously used raid disk before matting it up with your cloned disk.

Do I need to run chkdsk before creating the Acronis image?....I had an images fail to restore,once, with no great consequence due to OS level errors (still had the original array intact), an associate had the same, but did not have the original array, so it was a disaster..so I make sure I run chkdsk before imaging or cloning.

Active Directory deletes objects (users, servers etc) automatically within it's database after a period of time. Basically if your clone sits around past the tombstone life, your AD is toast if you use the clone, so you need to clone again before the tombstone life is exceeded. On my AD networks, I have set it's life to 180 day. The default is 60 days with Win 2000 and 2003, 180 days with 2008 , but you can modify Win 2000 and 2003 to go 180 days. Also be aware, with AD, if you need to use the cloned disk and you have another domain controller, there are replication issues. (Ah, life was so much simpler with work groups :emotion-3:)

http://support.microsoft.com/kb/216993

http://windowsitpro.com/article/articleid/41576/ad-tombstone-objects.html

http://technet.microsoft.com/en-us/library/cc782557.aspx

Lastly, you can pull the disk out with it running, and replace that disk with another, again while running

Thanks again for your reply. I havent been able to log in the last few days.

If I have a clone that is outside the tombstone period can I still not use it and then restore my system state backup?

I like the idea of clones but not having to do them too frequently. If i can set it to 180 days thats not too bad though.

When you've created a cloned disk where do you store this?

All I need is a solution that allows business continuity. RAID 1 on the system drive supports this until corruption replicates to the other leaving me having to restore.

The clone would support a restore but there is the AD tombstone issue (also, would this clone be bootable without having to do too much? What do you mean by low MTBF? Do you mean if I keep pulling out and replcing drives its likely to wear more quickly?)

I also will have an Acronis image, which i dont really trust but will keep as an extra backup.

Have you worked much with acronis because I'm lead to believe the acronis image if restored will not be bootable?