Koozali.org: home of the SME Server
Contribs.org Forums => General Discussion => Topic started by: bpkheops on August 28, 2014, 10:39:59 AM
-
Hi all,
I have seen that SME Server 9 is released in RC1 version.
Do you know if it supports hard drive more than 2 TB ? It's based on centos 5.6 but I don't know if centos 5.6 support this capacity and may-be centos 5.6 is not enough to support it , customizing of sme server scripts are may-be necessary?
I have read some topics on these subjects, it seems the OS need lots of customizing to support hard drive with GPT table... : grub2, parted, etc...
-
mmhh...
SME9 is last stable release
SME9 is based on CentOS 6
-
Perhaps a little more reading of the wiki would be of benefit
-
ok my mistake, I mean centOS6.
If I ask on forum, it's because there is no information on wiki for SME Server 9 about that.
I have found by search that SME Server 8 don't support disk > 2 TB, that's why I ask for SME Server 9.
-
I would say yes, of course..
in any case, you can check with google searching info for CentOS 6.X
-
bpkheops
I have found by search that SME Server 8 don't support disk > 2 TB, that's why I ask for SME Server 9.
It depends what you mean by support. IIRC boot partitions larger than 2Tb are not supported, but you can have data drives larger than that. Did you search the forums as there have been a number of posts about it eg search on gpt. Also search bugzilla as there have been discussions there too. You can add hardware controllers to support thousands of Tb's.
SME9.0 final release is based on CentOS 6.5 so search support for that version of CentOS.
-
Ok I have tried to install SME 9 on one 4 TB hard drive. The install scripts works, it will create 2 partitions, so the install works with GPT table, but you can't boot on this disk. I have grub but it can't find kernel. My motherboard has efi bios (prerequisite to support booting from hd > 2 TB).
May-be when Grub2 will be used, it will work. I am going to look around in bugzilla.
I try to use my 4 TB hd as system and data disk, not only data disk.
If I find a solution, I will update this thread ;-)
-
Fyi this howto exists....it is not as you want to do but it is a begining (http://wiki.contribs.org/Raid:Creating_large_raid5_array_(over_2TB_drive))
-
Two posts re CentOS 6 and drives >4tb.
Seems like it can be done BUT by a third party partioning tool like Gparted AND the mainboard must also support UEFI etc..
https://www.centos.org/forums/viewtopic.php?f=14&t=44367
http://grokbase.com/t/centos/centos/1457k5f9vp/format-big-drives-4tb-in-the-installer
-
In fact, I did not success to install because I used one USB key.
The USB key is seen like a HDD.
The install scripts copy the files from the key to hard drive BUT the grub installer write on USB key, not on the hard drive!!
So I used a CDROM and now I have sme9 online!!!
SME 9 scripts work with 4 tb HDD and you can boot on it.
Thank to all for your replies.
-
bpkheops
SME 9 scripts work with 4 tb HDD and you can boot on it.
Well that's good news.
Do you mean you just did a normal install from CD & did not run other drive management scripts etc.
So you now have a single 4Tb HDD with boot, system & data on the one 4Tb partition, is that correct ?
Please show us output of
df -h
&
cat /proc/mdstat
-
So you now have a single 4Tb HDD with boot, system & data on the one 4Tb partition, is that correct ?
correct.
I flag my hdd drive before as gpt table with parted, I don't know if it help to install.
[root@rangiroa ~]# cat /etc/e-smith-release
SME Server release 9.0
[root@rangiroa ~]# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/main-root
3,6T 1,3T 2,2T 37% /
tmpfs 1,8G 0 1,8G 0% /dev/shm
/dev/md0 243M 50M 181M 22% /boot
[root@rangiroa ~]# cat /proc/mdstat
Personalities : [raid1]
md0 : active raid1 sda1[0]
255936 blocks super 1.0 [2/1] [U_]
md1 : active raid1 sda2[0]
3906629440 blocks super 1.1 [2/1] [U_]
bitmap: 30/30 pages [120KB], 65536KB chunk
unused devices: <none>
-
Same results here, SME 9.1 installs fine on 4 TB hdd, creates 2 GPT partitions.
Warns about UEFI boot, but boots up ok (actually under Hyper-V on 2012R2),
and after messing around with parted (creating exactly same partitions as on
sda) managed to manually add and sync up the second physically mapped HDD.
Question is why doesn't manage disk redundancy from console handle correctly
GPT partitions. Is there anything that can be done, modify scripted action/event?
-
Question is why doesn't manage disk redundancy from console handle correctly
GPT partitions.
Because it uses sfdisk which doesn't know how to read or create GPT partition tables. It was written long ago.
Is there anything that can be done, modify scripted action/event?
The script could be re-written to use different tools (e.g. parted). You'd need to suggest that via bugzilla. But it'll only be implemented if someone wants to implement it, or if someone offers to pay to have it implemented.
-
http://bugs.contribs.org/show_bug.cgi?id=9182
Thanks.
-
Another alternative could be sgdisk (if it's available under RHEL)
http://www.cyber-tec.org/2012/04/07/sfdisk-for-gpt-we-use-sgdisk/
Copy the partition scheme from /dev/sda to /dev/sdb:
sgdisk -R=/dev/sdb /dev/sda
Now randomizes the GUID:
gdisk -G /dev/sdb
-
Another alternative could be sgdisk (if it's available under RHEL)
gdisk is available in the base repository for SME9. (yum install gdisk)
Best would be if you would add your comments to the bug instead of here. The devs will look at bugs rather then the forums.
-
I thought I would add to this thread as it is about disks > 2TB. I upgraded my server just over 2 years ago with 4 x 2TB HDD and I have run out of space again. I want to replace the 4 x 2TB HDD with 6 x 4TB HDD to take me from 5.5TB to 19TB of storage with it all as one volume and preferably using xfs.
I have spent the last week reading that many articles that I became confused, I was trying to find a way around fdisk and it's 2TB limit and after some experimenting I think I have a way to do this at the moment on a test PC that has an Asus H170 Pro MB, similar to the MB in my server. What follows is a bullet point description of what I have done.
- In BIOS enable UEFI and choose “Other OS”, 1 x 500GB and 1 x 1TB sata disks installed.
- Boot from the Centos 6.7 Live DVD.
- Install gdisk, create a GPT partition on both disks and then partition sda1 and sdb1 as 500M with a GUID of fd00 (Linux RAID). Then partition sda2 and sdb2 as 465G with a GUID of fd00 (Linux RAID).
- Run mdadm and create a RAID set md0 on sda1 and sdb1 with metadata=0.90 for booting. Then create a RAID set md1 on sda2 and sdb2 with default metadata.
- Create a pv on md1 and then create a vg called main.
- Boot off the SME Server 9.1 CD, select Advanced installation options and then select to install from the graphical interface.
- Select the normal options and then select Create Custom Layout for disks.
- Click on md0, tick the format box and set it up as /boot.
- Click on vg main, select edit and add lv root, mount point / of 30G, then add lv swap, mount point swap of 10G.
- Proceed with the installation, reboot and configure and then do an update.
- Install gdisk and xfsprogs.
- Create an lv called files in vg main using all the remaining space in main and format it with xfs.
- Make a backup of the directory structure below /home/e-smith/files and then delete all.
- Mount the new lv files using mount /dev/mapper/main-files -t xfs /home/e-smith/files.
- Make sure the lv is mounted and then copy the directory structure back to /home/e-smith/files.
- Edit /etc/fstab and add the entry for lv main-files then reboot and check that it has mounted correctly.
This is repeatable, I have done it twice so far with exactly the same result but I haven't done a lot of testing like creating users and testing access as a normal user which I will do. I am assuming that there should be no real difference between doing this on the small HDD I used and doing it on some 4TB HDD, can anybody think of a reason why this wouldn't work with larger disks.
[root@test ~]# cat /proc/mdstat
Personalities : [raid1]
md0 : active raid1 sda1[0] sdb1[1]
510912 blocks [2/2] [UU]
md1 : active raid1 sda2[0] sdb2[1]
486943744 blocks super 1.2 [2/2] [UU]
bitmap: 1/4 pages [4KB], 65536KB chunk
unused devices: <none>
[root@test ~]# mount
/dev/mapper/main-root on / type ext4 (rw,usrquota,grpquota)
proc on /proc type proc (rw)
sysfs on /sys type sysfs (rw)
devpts on /dev/pts type devpts (rw,gid=5,mode=620)
tmpfs on /dev/shm type tmpfs (rw)
/dev/md0 on /boot type ext4 (rw)
/dev/mapper/main-files on /home/e-smith/files type xfs (rw)
none on /proc/sys/fs/binfmt_misc type binfmt_misc (rw)
[root@test ~]# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/main-root
29G 1.8G 26G 7% /
tmpfs 3.8G 0 3.8G 0% /dev/shm
/dev/md0 476M 57M 394M 13% /boot
/dev/mapper/main-files
426G 33M 426G 1% /home/e-smith/files
[root@test ~]#
-
i'm looking for a new 4TB HDD,
can anyone help me to find the best one?
-
welcome here
my 2c: avoid "green" hds if you're thinking to use raid (and you're strongly encouraged to do so)
if you can, install SME on smaller disks then add the bigger ones following the howto you read above
that said, choose the one you prefer.. if you can, buy disks from different suppliers..
I have no experience with such disks, maybe someone else can help you more than me
-
I prefer Hitachi HDD currently because of their good track record and I used Hitachi 4TB NAS HDD in my server.
-
I found this article on google
http://www.deskdecode.com/top-best-4tb-hard-disk-drive-monthly-updated/ (http://www.deskdecode.com/top-best-4tb-hard-disk-drive-monthly-updated/)
and they suggest me to buy Seagate Barracuda.
how about that?
-
it's up to you ;-)