We are in the process of migrating this forum. A new space will be available soon. We are sorry for the inconvenience.

Reallocated sectors and high fly writes - disk failing, software RAID question


Andy
11-11-2013, 17:22
Thanks, but I have a friend in Sweden who has 1000/100 and can handle it all, but it still takes a good 5-7 days to upload it all again at best.

rickyday
11-11-2013, 17:20
Quote Originally Posted by Andy
It is, but they're just downloadable files from my website not immediately required data. Only about 30-40GB is website/database etc data.
I have 1.27TB free on my mSP if you need anywhere to transfer large files/ISOs to Andy?

Offer is there.

EDIT - Also 1TB on a 4G spare if that is needed.

Andy
11-11-2013, 15:48
It is, but they're just downloadable files from my website not immediately required data. Only about 30-40GB is website/database etc data.

yonatan
11-11-2013, 15:46
Quote Originally Posted by Andy
Downtime doesn't bother me that much, it's the fact that OVH's current setup doesn't allow me to keep all of my data on the remaining 2nd disk unless I have a RAID setup that works. 1.2TB is a lot to upload, let alone 2.5TB if both get wiped.
Yes, with the current setup it seems there is no way around it..
that 1.2TB is a lot for a single server which needs backups, .. well at least you got 5 days to copy at least 1TB of backups for free

Andy
11-11-2013, 15:37
Downtime doesn't bother me that much, it's the fact that OVH's current setup doesn't allow me to keep all of my data on the remaining 2nd disk unless I have a RAID setup that works. 1.2TB is a lot to upload, let alone 2.5TB if both get wiped.

yonatan
11-11-2013, 15:35
Quote Originally Posted by Andy
I don't, that's the issue. This seems to only apply to hardware RAID
well, if it's matter a of priority i suggest you implement some DRP to avoid this issue in the future.

Since you are running primarily on Windows,
a true minimum downtime solution would require two servers and to play a bit with hyper-v backup/migration system.
http://www.petri.co.il/move-virtual-...r-v-hosts.htm#

you could also trigger an IP Failover switch with some fancy API commands & scripting.


for a one server setup, just always keep good local backups.
softraid is only one layer of protection, normally when softraid fails , that means a format on a headless setup.

a great way i found to backup a single windows server was to combine the windows backup system with the NAS service

http://www.ovh.co.uk/nas/
+
http://technet.microsoft.com/en-us/l...(v=ws.10).aspx

if you are currently able to access the disk from rescue, you can mount the NAS and copy your data manually over and re-image right away.

Andy
08-11-2013, 19:39
I don't, that's the issue. This seems to only apply to hardware RAID

yonatan
08-11-2013, 19:27
provided you have kvm access to the machine, when it boots up , you should have a selection to boot from " Secondary Plex ".
the boot order must be changed from the bios to let it pick up the second disk correctly.

Andy
08-11-2013, 14:24
That's what I mean, the boot code isn't within the main partition so software RAID1 will never work without some modifications from what I can tell. Only H/W RAID1 will allow it to work seamlessly.

NeddySeagoon
08-11-2013, 13:39
Andy,

boot.ini is in the filesystem, so its covered by the raid1.
The problem/risk is with the boot code outside of the filesystem.

Andy
08-11-2013, 13:21
The problem is if the boot.ini is only on the first disk and that drive is gone (removed for replacement) how will it boot? There is no boot loader anymore. I thought the whole idea of RAID1 was that the drives were mirrored exactly as the other one is. The only thing that would be determining where it boots from is the BIOS...

My tests have shown this doesn't happen. I installed a virtual machine last night with software RAID1 and removed the first disk. All I got was an OS not found error from the BIOS.

The main reason I'm asking this is because I have 2.5TB of data on this server and whenever OVH replace the boot drive the re-install system forces the wiping of both disks which is plain stupid. I have nowhere to backup 2.5TB of data and re-uploading takes weeks.

DigitalDaz
08-11-2013, 02:25
Many years since I did all my MCSE stuff, back in the 2K3 days.

If I remember rightly you had to add another line to the boot.ini file, this would allow you to choose which disk to boot from in a software raid one.

Just done a quick google and I definitely remember this:

[boot loader]
timeout=30
default=multi(0)disk(0)rdisk(0)partition(1)\WINDOW S
[operating systems]
multi(0)disk(0)rdisk(0)partition(1)\WINDOWS="Micro soft Windows Server 2003" /noexecute=optin /fastdetect
multi(0)disk(0)rdisk(1)partition(1)\WINDOWS="Micro soft Windows Server 2003" /noexecute=optin /fastdetect

After that I saw the light, abandoned MS and went Linux so I don't know about 2K8 onwards.

JakeMS
07-11-2013, 23:12
In my experience it would still work after a reboot with a new disk, even if it is the primary disk.

But to do so, you need to ensure:

1) Your bootloader is installed on BOTH disks
2) The BIOS knows to switch to the next drive if it can't boot off of one (This is down to OVH to configure in this case)
3) Your software RAID is configured correctly.

So I think you'll be okay as most of the above is already done . If not, just quickly install grub on your secondary disk and hope for the best .

Andy
07-11-2013, 20:15
The problem I'm having is that if they pull the original boot disk, in software RAID1 you might need to do some more config than just putting a new disk in as the boot loader is on the first disk not the 2nd. Reading up online hasn't revealed much information.

Kode
07-11-2013, 19:55
Quote Originally Posted by Andy
I've had disks changed before within 24 hours of the request. Certainly not quick considering the SLA I'm supposed to have on my server but not a silly amount of time.
Isn't the issue more a case of trusting them to pull the right disk? I'm sure I read somewhere on this forum about them pulling the wrong disk and everything getting lost.

ctype_alnum
07-11-2013, 18:19
I am not 100% certain but if a new disk is added you would go to disk management and add that to the existing array then it would begin to populate the new disk etc. I did have a windows 7 machine with two drives running in raid 1 for a few months and it was pretty decent.

Andy
07-11-2013, 17:18
I've had disks changed before within 24 hours of the request. Certainly not quick considering the SLA I'm supposed to have on my server but not a silly amount of time.

macole111
07-11-2013, 17:17
Good luck getting OVH to change a drive, that is one of my biggest fears having a critical disk fail.

-macole111

Andy
07-11-2013, 14:29
I've got possible issues with a failing disk again, the one that got replaced just months ago (it's Seagate, what was I expecting?). It has 7 reallocated sectors and 250+ high fly writes which indicates to me a defective/failing drive.

My question is this:

If I am using Windows software RAID1 for my boot partition between the two disks, can OVH replace the first disk and then the RAID1 on the other drive take over for booting the server? I've never tried a Windows RAID1 software setup for the boot drive before so I don't know if this will work without additional input.

Thanks.