AJWhite1970 Posted September 9, 2014 Posted September 9, 2014 Hi Been phoned by a primary school in a panic this morning as they have come in to find a Dell server with two failed drives in a RAID 5 configuration. It runs their Exchange 2003(!) server but no home drives etc, that's on the other server. The Raid manager is showing:- 00: Ready 01: Failed 02: Online 03: Online 04: Online 05: Failed The server obviously won't boot. My understanding of Raid 5 is if you loose 2 drives at once it can't be recovered even if you put in new disks? Am I correct in telling them it's unrecoverable? Thanks Andrew
tmcd35 Posted September 9, 2014 Posted September 9, 2014 Were the drives definitely in the same array? I only ask because the more typical config is Drive 00 and Drive 01 mirrored and drives 02-05 in RAID-5. In that scenario the two failed drives are on different arrays and they may yet rebuild with fresh drives. On the other hand if the drives are on the same array then yep, I'm afraid all data is lost. If you get the drives to a professional forensic recovery outfit, and keep them in the right order, they might be able to get some data back (££££'s). Backups? 1
AJWhite1970 Posted September 9, 2014 Author Posted September 9, 2014 Backups? Don't think so from the way they are panicking..... How would they be able to check what was in which array? Is that in the RAID config somewhere? I'm trying to do this by phone and screen shots before I get in the car and drive 80 miles.. Cheers Andrew
tmcd35 Posted September 9, 2014 Posted September 9, 2014 It'd certainly be in the RAID BIOS config somewhere. Problem is it's different for every controller. Usually they have a view config option on the main menu and from there it should be reasonably obvious if it's one array configured or multiple. Most primary school servers I've seen have had 2 drives mirrored for the OS and the rest RAID-5, but there are exceptions to every rule. If it was set up with a separate OS array you'd have expected the machine to boot, but then some controllers do refuse to boot in a degregated state to avoid further chance of data loss. 1
Rodan Posted September 9, 2014 Posted September 9, 2014 If you’re desperate it's always worth reseating the drives in the bays. Sometime drives are failed due to a bad write threshold being crossed and reseating them can clear this. Not guaranteed but worth a go if all else is lost. 1
AJWhite1970 Posted September 9, 2014 Author Posted September 9, 2014 Got them to reseat the drives and then they decided to force both drives online in the Raid controller despite my protests. Amazingly the server has rebooted with all data present. After I've given them the lecture about data backups etc, I'll start getting them to take a vhd snapshot every night so at least there is recoverable data. Hopefully this scare is the kick up the ar$e they need. Thanks for all the advice Andrew
Joanne Posted September 9, 2014 Posted September 9, 2014 Might be worth replacing the troublesome drives anyway just incase.
AJWhite1970 Posted September 9, 2014 Author Posted September 9, 2014 It's an old 2005 Dell server with "strange sockets on the back of the drives" which I'm guessing is the old SCSI connectors. Not sure sourcing a replacement for that would be that that easy. Am I right in thinking any larger drive is OK or has it got to be an exact 300Gb replacement? Either way not my problem but I bet I get asked.... Andrew
localzuk Posted September 9, 2014 Posted September 9, 2014 My suggestion would be that if they are able to run it on a server that old, a new replacement server of suitable spec wouldn't be a huge cost... However, I'd also suggest updating to something from this decade software-wise too! They're very lucky that they could force it back online!! 1
3s-gtech Posted September 9, 2014 Posted September 9, 2014 Ultra 320 SCSI drives aren't hard to get hold of still. You can happily use a larger drive, it won't need to match exactly. Try to make sure it has the same spindle speed though. 1
john Posted September 9, 2014 Posted September 9, 2014 Also Exchange 2003 is EOL and has been for a bit so should be considered a large security risk really now, especially given its role it will be exposed to the internet - Save The Date - End of Exchange 2003 Support–T Minus 6 Months - 250 Hello - Site Home - TechNet Blogs So time they modernise and upgrade a bit as well 1
AJWhite1970 Posted September 9, 2014 Author Posted September 9, 2014 I agree with all your comments, just arrived on-site and it's a museum here rather than an IT lab. I've brought up one of my older development servers and I'm going to virtulize the exchange server as soon as school finishes and then discuss proper upgrade plans. Windows 2003 with Exchange 2003, the other server here is also 2003 and struggling for disk space. I think this could turn into one of those phone calls I wish I'd never taken... Thanks everyone for your comments Andrew
RobD Posted September 9, 2014 Posted September 9, 2014 Could try and reseat the dead drives, might be lucky.
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now