opennode 6.0rc1 install failed

[chenlong828]

I was trying to install the opennode 6.0 rc1 to my IBM HS22 Blade Server though the Remote Control of Advanced Management module.

When I finished the text install step and got into the GUI installer, I encountered my problem:

After assigned the root password, the install tell me that "Unable to read group information from repositories. This is a problem with the generation of your install tree."

The dialog with only two button: Exit installer and retry(without help)

Before enter the GUI installer, I choose automatic raid 1 with softraid as my partition profile.

What can I do to continue my experience?

[activesys]

Please try also install without softraid - and see if it works. We will try to replicate the problem but it might be hardware specific - eg how drive devices are detected and assigned. You can convert later to softraid on-line (without stopping or re-installing the server) - I can provide you with exact instructions for that later if you want.

[activesys]

We are unable to reproduce the problem in VirtualBox environment for softraid install - eg install proceeds normally beoynd password dialogue - so its probably hardware specific. Could you please also choose install with manual partitioning and see how blade disks are represented in manual partitioning dialogue - and please post us the findings - as we dont have currently HS22 blades in our hands

[activesys]

Also - have you checked your install ISO md5 checksum - in order to verify that downloaded ISO file is not corrupted in any way?

[activesys]

I added softraid conversion howto if anybody interested: http://opennode.activesys.org/forum/index.php/topic,158.0.html

[chenlong828]

I have checked the md5sum, it's the same as sourgeface, so my ISO file is correct.

I choosed the manual partition and the automatic partition, but all without help, it appear the same mesage again.

The installer can find my 2x146GB scsi harddisk, and form slash message appeard in the screen, the lvm partiton process is passed.

What is the install tree? Does there have some method to output some log?

[chenlong828]

I choose the manually in the text mode and in the GUI I choose use all disk, checked and review the disk map:

LVS group:
lv_root with 50GB
lv_home with 190GB
lv_swap with 32GB

the hard disk has been partitioned like this:
sda: 146GB
0.5GB for boot
the left to lvm partition(VolGroup)

sdb:146GB
all of the space to lvm partition(VolGroup)

[activesys]

Install tree is Packages folder on ISO - so the error msg says that access to it is either corrupt or files are corrupt, etc. But if it would be really corrupt on released ISO - then there would be more complaints and it would not work in tests also - but it does. So the issue seems to be either hardware specific or tied particulary to your case/setup. There is more options you can do:

1) Try earlier OpenNode 6.0 beta1 ISO installer - see if it makes any difference - you can issue yum update later and get the fully updated system anyway

2) Try original CentOS 6 installer and convert the system to OpenNode later using the centos6-to-opennode6 conversion script. Please report any script errors back - as it might need some cosmetic changes - thou it should work.

[chenlong828]

Could I upgrade from Centos 6.2 or just CentOS 6.0?

[activesys]

CentOS 6.2 should be fine - install from minimal CentOS 6.2 ISO.

We usually partition OpenNode as:

/boot 1GB (ext3)
rest of disk as LVM PV
on LVM
swap 2xRAM but not more than 8G
/ min. 5G
/storage 5-25G initially - later can be expanded online
/vz  5-25G initially -  later can be expanded online

Leave at least 15G free space in LVM VolGroup - for LVM snapshots

[chenlong828]

I have successfully finished the convert process.

During the process, I got these message:

Failed to create a new pool: Failed to execute command 'virsh 'pool-define-as local dir --target /storage/local''. Status: '256'. Output: 'error: Failed to reconnect to the hypervisor
error: no valid connection
error: Failed to connect socket to '/var/run/libvirt/libvirt-sock': No such file or directory
'
  Installing : opennode-6.0-10.x86_64                                                                                 213/217
error: "net.bridge.bridge-nf-call-ip6tables" is an unknown key
error: "net.bridge.bridge-nf-call-iptables" is an unknown key
error: "net.bridge.bridge-nf-call-arptables" is an unknown key
error: "net.ipv4.conf.vmbr0.proxy_arp" is an unknown key
set forward delay failed: No such device

Does it correct?

And the another question, can I use another card for bridge network?

thanks.

[chenlong828]

I was starting download the OMS image, and later I will try and report it^^

[activesys]

1) These errors are ok - as there is no existing libvirt and bridge - so online modifications fail but they are still configured to be present after reboot

2) You can use whatever network interfaces - just reconfigure vmbr0 bridge members - eg take eth0 out and put something else in.
See brctl help for syntax - in order to manipulate bridge member devices.