JP Vossen on 16 Jun 2011 00:15:39 -0700

[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

[PLUG] 5TB RAID5, GTP & CentOS-5: what's going to bite me?

I posted about this a while ago, but I finally got CentOS-5.6 to boot from a 5TB hardware RAID5 device, using a GTP partition table and Grub2.

Normally, CentOS-5 will only boot from a device smaller than 2TiB due to MBR limits [1]. Using a GPT partition table fixes that, but Grub Legacy doesn't know about GPT; you either need a patch or Grub2. Ubuntu switched to Grub2 in 9.10, by the way...

As an aside, one of the things I found out is that RHEL6 is *still* using Grub Legacy [2], though per the spec file change log it was patched for GTP support in 2007. So, Grub1? WTH?!? And patched in 2007? So why not RHEL5 GTP support?!? AAAARRRGGHHHHH!!!

I tried using that RHEL6 patched Grub1, but it had build-deps that CentOS-5 couldn't satisfy, so I went the other way and looked at Fedora and FC. I ended up with FC10 because RHEL5/CentOS-5 are circa 2007 and FC10 was the oldest "FC" that had Grub2. So I downloaded the FC10 Grub2 SRPM and rebuilt it on a CentOS-5 machine after installing a few build dependencies, then:
1) Configured the hardware RAID5
2) Booted an Ubuntu 10.04 LiveCD to use gparted to do the GPT and partitions
3) Rsynced my CentOS-5 install from a similar server
4) Booted a CentOS-5.5 LiveCD (OS is 5.6, but the 5.5 LiveCD was handy)
6) Mounted and chrooted into the root on the RAID5
7) grub2-install /dev/sda2
8) Rebooted to the RAID5

It booted, and it seems like everything more-or-less works. fdisk/sfdisk complains they don't know about GTP and say to use gparted, otherwise I haven't found any problems. Yet.

What haven't I thought of that's going to bite me?


"...the maximum size of a partition or the maximum start address (both in bytes) cannot exceed 2.19 TB"

JP Vossen, CISSP            |:::======|
My Account, My Opinions     |=========|
"Microsoft Tax" = the additional hardware & yearly fees for the add-on
software required to protect Windows from its own poorly designed and
implemented self, while the overhead incidentally flattens Moore's Law.
Philadelphia Linux Users Group         --
Announcements -
General Discussion  --