Fresh Install UCS 4 "Loading, please wait"

After several years of working with ClearOs I was pinpointed at UCS.
Which look promising :slight_smile: My old home server is at the end of life time.
So I bought an new Dell T20 with 2 x 3TB WD disks.

The initial installation goes fine disks can be partionized (GPT) and put into RAID1.
Partition Table:
1 MB Reserved BIOS boot area
1 GB /Boot (Ext4 On Raid 1)
5 GB Swap (Each disk has 1 swap partition so in total 10GB swap)
3 TB / (Btrfs on Raid 1)

At the moment the system should reboot it shows the Grub 2 with 4 options
1 Kernel from disk 1
2 Recovery from disk 1
3 Kernel from disk 2
4 Recovery from disk 2

After 3 seconds the UCS boot logo is shown on the screen and then…
The message “Loading, please wait”

Were it sticks.

Where did I go wrong??

Hi,

it might be of help if you press “ESC” during boot (when the splash screen shows up). You should see the boot messages than which hopefully reveal the problem.

Just one note: You should put your swap partition on RAID too. If you have both swap partitions active and one disk fails, you might loose data in swap.

Regards,
Janis Meybohm

Hello
I indicates the following error message:

(process: 311): GLIB-warning getpwr_id(): failed due to unknown user id ()
mdadm: devices UID-xxxxxxxx:xxxxxxxxxx and UID-xxxxxxxx:xxxxxxxxxx have the same name /dev/md/1
mdadm: duplicate ID device names in conf were found 

Hello,

this does not look good.
I just tried to reproduce your setup and it worked fine for me (UCS 4.0-1).

I’ve created the partitions as you described:

Then used the “Configure software RAID” option to setup the two md’s:

After installation (and reboot) both raids are active with their unique id’s:

Could you please verify if you are able to reproduce this?
You could also try to boot up live system, assemble the raids, mount your / and /boot partition, fix the /etc/mdadm/mdadm.conf and run a “update-initramfs -kall -u” in chroot.

Regards,
Janis Meybohm

I have tried to reinstall UCS several time.
But some how I creates every time the same ID.

When I try to open the drives via an live linux version it is also generating errors.

Could not stat device /dev/md/0 - No such file or directory
Could not stat device /dev/md/1 - No such file or directory

During the re installation I discovered that the MD’s were disappeared?

Further more I can see different names then expected for the raid arrays.
They are called md126 = 3TB and md127 = 1GB.
Within the /etc/mdadm/mdadm.conf
there are 2 MD1 references:

ARRAY /dev/md/1 metadata=1.2 UUID=xxxxx:xxxxxxx:xxxxxxx:xxxxxx name=unassigned-hostname:1 ARRAY /dev/md/1 metadata=1.2 UUID=yyyyyy:yyyyyy:yyyyyy:yyyyyyname=unassigned-hostname:1
Only the second is matching to the disks.

I tried to disable one line also this didn’t worked.

I just noticed I had the “old” ISO UCS 4 I am downloading - burning etc 4.01 perhaps this solves it…

Hello,

[quote=“2Psycho”]When I try to open the drives via an live linux version it is also generating errors.

Could not stat device /dev/md/0 - No such file or directory Could not stat device /dev/md/1 - No such file or directory [/quote]Device names may be different, depending on the linux distribution and kernel version used. You will also have to assemble the arrays manually before the divices will appear (check /proc/mdstat).

[quote=“2Psycho”]Further more I can see different names then expected for the raid arrays.
They are called md126 = 3TB and md127 = 1GB.
Within the /etc/mdadm/mdadm.conf
there are 2 MD1 references:

ARRAY /dev/md/1 metadata=1.2 UUID=xxxxx:xxxxxxx:xxxxxxx:xxxxxx name=unassigned-hostname:1 ARRAY /dev/md/1 metadata=1.2 UUID=yyyyyy:yyyyyy:yyyyyy:yyyyyyname=unassigned-hostname:1
Only the second is matching to the disks.

I tried to disable one line also this didn’t worked.[/quote]Where do you get those names from (md126, md127)? And what do you mean by “is matching to the disks”?
The first parameter of the ARRAY line (device name) should be different (see my last screenshot) but you will need both lines for the system to boot.
Please also check the output of “mdadm --detail --scan” and “blkid”.

Regards,
Janis Meybohm

I got the names from the live distribution Parted Magic.

I also found a few other strange things there for I decided to complete reformat (low level) both HD’s.
After this job I was able to install the UCS in Raid 1.

So Far So Good :slight_smile:

Mastodon