Koozali.org: home of the SME Server

Obsolete Releases => SME Server 9.x => Topic started by: Brave Dave on November 22, 2014, 03:12:00 AM

Title: Bootloader / MBR - Large Disks
Post by: Brave Dave on November 22, 2014, 03:12:00 AM
This statement might be mixed up - apologies in advance if the  concept is skewed

With larger Disks and with SME 9 (or rather CentOS I guess), the disks are using GPT rather than MBR

So if I take my RAID 5 with 3 x 3TB disks, and add an additional disk as a Hot Spare (that is, SME8 style)

Between these howtos the information seems relevant

And I get my additional disk into the array as I require

My Question is - is this still a correct procedure for creating the boot record on the larger disk ?
 ( the hidden question is Will it boot off this drive ? )

Title: Re: Bootloader / MBR - Large Disks
Post by: janet on November 22, 2014, 08:58:24 PM
Brave Dave

http://forums.contribs.org/index.php/topic,51075.msg258621.html#msg258621
Title: Re: Bootloader / MBR - Large Disks
Post by: stephdl on November 22, 2014, 09:27:48 PM
Thanks Janet to bring us that post I had forgotten....well it is up to you to test directly brave dave and report here if you can install directly the sme9 on your four drives.

If you see the pointer, bpkheops has flagged the hdd drive before as gpt table with parted, but he doesn't know if it help to install.

Please try and report here
Title: Re: Bootloader / MBR - Large Disks
Post by: Brave Dave on November 23, 2014, 07:16:41 AM
Quote from: janet on November 22, 2014, 08:58:24 PM
Brave Dave

http://forums.contribs.org/index.php/topic,51075.msg258621.html#msg258621

Thanks for the pointer.

My Understanding is that they used sme to install a new machine and it boots - and I agree with that

My test here is that a disk can fail, and I can continue. SME8's RAID could tolerate this, but I found a .. unreliability.. unless I wrote the bootloader manually using grub

I can't actually verify this because I just crashed the system, I'll try this again

This is a new system, I'm just trying to understand what's going on. I'll update the forum when the system is rebuilt and I've pulled a disk.
For the record

I don't know how to build SME 9 with a hot spare, it builds it with 75% disk space (i.e. 4 x 3TB disks deliver 9TB of space, SME 8 would deliver 50%). So I build on 3 disks and add the spare manually ..

to my thinking - pulling the disk is the same as a disk failure. SME8 Raid5 would send an email (you could actually pull 2 disks if you knew which ones to pull), rebuild the array using the hot spare and continue, and I think SME9 will too, but I failed to test this successfully so far, and I think it is because the boot record isn't written to the subsequent disks - perhaps