Mirrored disks - how to replace?

[diagram]

I have been having some problems (lots) with a pair of mirrored disks. These hold our user folders. I replaced one of the disks, and still have a few messages "at risk" etc.

 

Is there an easy way I can copy off the information and user shares etc to another set of mirrored disks? (It is hard to get non-Advanced Format disks now, I would like to be able to set up on two new Advanced Format disks.

 

The current mirror is of course a dynamic volume which doesn't make things any easier.

[Michael]

If you've only replaced one of the disks, it could well be the other disk also needs replacing.

 

Of course it's always recommended to either move files to another server or computer and to have some form of backup when 'messing' with RAID arrays.

[diagram]

I'm a bit nervous about restore, never having used it in earnest before.

 

If I restore to a new disk, will the Server 2008 Backup utility restore not only the files and folders but the shares, security info etc. as well?

 

At what point do I logically remove the failed/failing disks from the system?

[Michael]

The crucial part is the System State as well as the System drive, i.e. where Windows lives.

 

Generally speaking everything should be restored, but it's difficult to tell without knowing what software you use to backup.

 

As for replacing your Raided disks, again it depends very much on what you're using - for example software raid vs hardware raid, such as Intel Matrix.

[RobBaxter]

Firstly, @Michael is wise ... copy EVERYTHING off

 

If you have any form of compitent Hardware Raid, you should just be able to pull 1 drive and replace it with another drive. (I will assume you are aware of the requirements for Raid Setups dealing with drives / specs and sizes). The raid controller at boot will detect the serial on the drive has changed and ask you what to do. (Rebuild Array) and wait what unholy time it takes and make god damn sure you have a UPS in place.

 

If both drives are playing up, repeat the step with "the other drive".

 

Also in my experience, when both drives or (all drives) appear to be having issues, it's almost allways a issue with the controller it's self. Again... if you have a decent Controller, it should dump this out to your syslog (check it). I had a new server once that when under any kind of load would crash. The problem was with the raid controller's Cache Memory resulting in errors which caused IO Wait the system could not recover from effectively "Hanging the OS".

 

Good luck with your issue buddy, hope you get it sorted out

[diagram]

Ok thanks for all this.

 

My system is very simple (unfortunately not simple enough). I am using the built in facilities in Microsoft Windows Server 2008 for mirroring and backup/restore.

 

In the end I jettisoned all my old disks and restored my boot drive and user drive onto two new

Advanced Format (AF512e) disks using my Server 2008 install DVD and my portable backup disks.

I think I will rely on backup disks and not bother mirroring

[Michael]

Ok thanks for all this.

 

My system is very simple (unfortunately not simple enough). I am using the built in facilities in Microsoft Windows Server 2008 for mirroring and backup/restore.

 

In the end I jettisoned all my old disks and restored my boot drive and user drive onto two new

Advanced Format (AF512e) disks using my Server 2008 install DVD and my portable backup disks.

I think I will rely on backup disks and not bother mirroring

 

So you're using software RAID, which is OK, but you can't beat a dedicated RAID controller of some kind. Most types of RAID are suitable to keep your system 'live', so in the case of a server, you can replace the faulty drive quickly and easily.

 

It's still better to have a combination of RAID1 or RAID5 (in most instances) for data redundancy as well as a backup. The odds of you losing both are pretty slim.