Lenovo/IBM ThinkServers

[rctempire]

Hello there.
We have a customer who has a Lenovo server. Apart from it essentially destroying its own raid, its own drives, memory, and a motherboard in one go. We have had to revive it and make it work.

So customer sits on SBS 2008 R2. Simple enough

4 SATA drives.
Raid 1 (2 and 2). Simple.
DISK 0 + 1 (RAID 1) - Windows SBS - Dying slowly
DISK 2 + 3 (RAID 1) - Company Data - Safe


Now two drives (Raid 0) one is dead. One of them is dying after it accesses a bad sector (about half way into the disk we think). It boots to windows and runs fine for hours, then drops off each time the bad sector must be accessed.

Ripped off the important SQL databases and im ready to rip exchange and AD off the machine in case it really goes wrong. Anything else ive failed to think of?


I have tried clonezilla and it fails at around 60% of copying so right now we are binning all the crap off the drive we don't need / WSUS crap.

Hopefully can clone this to another drive when its down to around 70GB.

Anyone got any tools for copying a drive with serious issues, we have sat it on a cool surface and with fans and keep blowing compressed air onto platters to keep it its temp low right now.


Also im now looking for ThinkServer EasyManage but struggling to locate an actual download of this. Anyone help?
EDIT - Ok found it http://support.lenovo.com/en_GB/downloads/detail.page?DocID=DS027544

Anyone for the clone part now?

 

[Longy]

Is it Hardware RAID or Windows RAID?

Where is the daily backups?

Just remove the faulty drive from the RAID1 - obviously you need to know which disk is bad - are they seagate drives?

Add a new disk and rebuild the mirror, I would then replace the other old disk and rebuild again at some stage.

If the server is having other hardware problems, I would get a new server with Hyper-V RAID10 and P2V the old server and run as VM.

 

[rdoxenham]

That's the problem with RAID 0 - you lose one disk, you lose them all.

Any chance you can boot to Linux and 'dd' the entire block device? That way you may be able to reconstruct the vast majority of the data.

 

[Longy]

Sorry just re-read that one disk is dead and the other one is showing problems on the OS RAID1.

Try a SBS backup to USB (this will make it easy to reinstall from the 2008 disc)

or

Try P2V of the whole machine and go Hyper-V

 

[Gareth]

I cringed when I read the title tbh, hate Lenovo / IBM servers, why anyone would buy anything other than HP is beyond me. Windows backup works surprisingly well in my experience, might be worth completing a backup to external drive then rebuilding and restoring once you've sorted the hardware out.

 

[rctempire]

Is it Hardware RAID or Windows RAID?

Where is the daily backups?

Just remove the faulty drive from the RAID1 - obviously you need to know which disk is bad - are they seagate drives?

Add a new disk and rebuild the mirror, I would then replace the other old disk and rebuild again at some stage.

If the server is having other hardware problems, I would get a new server with Hyper-V RAID10 and P2V the old server and run as VM.

Daily backups are good and safe.
Hardware raid - However on the MOBO it looks like if we power the SAS off, the memory goes with it, doesn't look to have a backup battery at all on it like DELL/HP.

That's the problem with RAID 0 - you lose one disk, you lose them all.

Any chance you can boot to Linux and 'dd' the entire block device? That way you may be able to reconstruct the vast majority of the data.

Didn't mean 0 - Its all raid 1.

Looking like we will do this.

Sorry just re-read that one disk is dead and the other one is showing problems on the OS RAID1.

Try a SBS backup to USB (this will make it easy to reinstall from the 2008 disc)

or

Try P2V of the whole machine and go Hyper-V

Trying to Hyper-V it, but even SSR is crapping out and killing the box saying cannot due to Bad sectors, even when the program said ignore bad sectors.

I cringed when I read the title tbh, hate Lenovo / IBM servers, why anyone would buy anything other than HP is beyond me. Windows backup works surprisingly well in my experience, might be worth completing a backup to external drive then rebuilding and restoring once you've sorted the hardware out.

I would totally agree with you, HP/Dell/Fuji for sure.
But we had Backup Exec on it and tapes are valid and verified, so may be using them tonight.

Just was after some ideas if any bright spark had thought of anything I hadn't.

 

[Cliotoby]

Whats wrong with IBM's ?! We have 800+ servers in our company across the globe and while over the last 3 years most of them have been virtualised the remaining ones and ones which need to remain physical are mostly IBM now, one shat its dimm last year but other than that they are faultless. All new hardware that goes into datacenter racks is IBM too - SAN's, Flex and Servers. Far as I know the people who look after the printers and desktops still buy HP but thats up to them.

 

[ChrisR]

Not even IBM wants their servers anymore, will have shiny new Lenovo labels on them now