0

I am wondering if someone can provide some help. I am trying to do an automated install (with kickstart file) with CentOS 7 on a CentOS 6 server. However I have failed multiple times so far. I did Google on it, but no such luck...

The latest source, which explains exactly what I am trying to do is located here.

So what did I do? I first downloaded the CentOS 7 vmlinuz and initrd.img files to /boot/

curl -o /boot/vmlinuz http://mirror.zetup.net/CentOS/7/os/x86_64/isolinux/vmlinuz
curl -o /boot/initrd.img http://mirror.zetup.net/CentOS/7/os/x86_64/isolinux/initrd.img

After that I edited the grub.conf with "vi /boot/grub/grub.conf" and added:

title Install CentOS 7
kernel /vmlinuz ks=http://my-server-ip/installation.ks
initrd /initrd.img

Sidenote: yes, I tested the IP from the server with the current CentOS 6 version and I could download/open it.

So in the end my grub.conf looked like this:

default=0
timeout=5
splashimage=(hd0,0)/boot/grub/splash.xpm.gz
hiddenmenu
title Install CentOS 7
        kernel /vmlinuz http://my-server-ip/installation.ks
        initrd /initrd.img
[older entries below]

and I also tried:

default=0
timeout=5
splashimage=(hd0,0)/boot/grub/splash.xpm.gz
hiddenmenu
title Install CentOS 7
        kernel /boot/vmlinuz http://my-server-ip/installation.ks
        initrd /boot/initrd.img
[older entries below]

Both ways fail for some reason. I had to physically check the server and I saw either of the following errors:

Error 1:

Error 15: File not found
Press any key to continue...

Error 2:

Error 1: Filename must be either an absolute pathname or blocklist
Press any key to continue...

Sorry, but I don't recall which error I received on which approach anymore.

There is one small / minor change to everything above. Instead of using CentOS 7 vmlinuz and initrd.img files, I used the ones from the OpenVZ 7 iso. Both files were located under the folder: /images/pxeboot/

The reason for this is that I want to install OpenVZ 7 on the server (so removing the current CentOS 6 version and replacing it with a clean, but kickstarted OpenVZ 7 version). I doubt the vmlinuz and initrd.img files are that different, so that it results in the above errors, right?

The kickstart file I created works. This was tested with PXE boot (on several different servers), however I want to be able to also install CentOS 7 or OpenVZ 7 (preferred) through the GRUB loader. This should be possible right?

Maybe I missed something, somewhere? Maybe someone has an idea what I am doing wrong or what I can try?

In the meantime I will retry again, but use the files from CentOS 7. I doubt that's the issue, but worth a shot anywas....

My test server specifications (dunno if it's important):

  • LSI Logic / Symbios Logic SAS2008 PCI-Express Fusion-MPT SAS-2 (RAID controller)
  • 2x HDD's in RAID 1
  • Running currently on CentOS 6.10 (Final)

The below might be related (sorry for the complete mess in this post, I am trying to be as detailed as possible). While I was typing the above server specifications I noticed a "similar" issue with a PXE boot machine with the same RAID controller card. Maybe that's the issue? Since I was fed up with failing the installation of OpenVZ 7, I decided to an automated kickstart with PXE (which worked on other serves). Mind you, this is another / different server, which has the same RAID card, however it was flashed to IT-mode and only has one (older) 400GB SAS SSD from Toshiba installed.

Anyways, running the automated installation (through the kickstart file) I noticed the server was hanging after I connected a monitor to it with the following (last) errors:

dracut-initqueue[723]: Warning: dracut-initqueue timeout - starting timeout scripts
dracut-initqueue[723]: Warning: dracut-initqueue timeout - starting timeout scripts
dracut-initqueue[723]: Warning: dracut-initqueue timeout - starting timeout scripts
x many more and followed by:
dracut-initqueue[723]: Warning: Could not boot
dracut-initqueue[723]: Warning: /dev/root does not exist

At this point I was like okay. You know what? I will do a manual installation of OpenVZ 7 through the PXE boot (I have an option for that).

So rebooted the server again, waited till it presented me the GRUB 2 menu (through the PXE server) and selected the manual install. I reached the menu and wanted to select the disk to install OpenVZ 7 on it. Now to my "surprise" it didn't find any drives!? So it could not continue the installation obviously. No matter how often I clicked "Rescan", no drives... At all...!

I have no clue what the issue is. But maybe it's the RAID controller being unsupported in CentOS 7 / OpenVZ 7? This one was flashed (orignally a PERC H200 Dell RAID controller, which equals a LSI-9211-8i) to IT-mode with a single Toshiba 400 GB SAS SSD connected to it (which is been seen by the controller during posting). So maybe this is the whole issue? Maybe someone knows?

I didn't notice the same controller until I typed server specifications above. Though one is IT-mode and the other one is in regular RAID 1 mode. So I tempted to think it's the controller, but then again the LSI-9211-8i is widely supported by all OS and versions, as far as I can tell. Even the current CentOS 6.x installation on it has no issues with it apparently?

Thanks in advance. And sorry for the big / detailed post.

Regards

HudsonHawk
  • 103
  • 2
  • 14

1 Answers1

0

Your grub.conf doesn't specify a root so it has no idea from which drive or partition to load the files vmlinuz and initrd.img.

You are also missing inst.ks= in front of the URL.

Instead of trying to do a hard drive boot, boot from CentOS 7 installation media and specify inst.ks=<URL> on the boot command line.

Michael Hampton
  • 237,123
  • 42
  • 477
  • 940
  • Thank you for taking tht time to answer. But if I want to install remotely (without physical access to the server) can I still install CentOS 7 on this server? For example load the CentOS 7.x installation remotely? That's what I am trying to achieve. In the provided link it shows it's possible? But I am failing, no matter what I try? – HudsonHawk May 27 '19 at 07:23
  • *" install remotely (without physical access to the server)"* - that's why most enterprise server systems come equipped with a built in an [out-of-band management](https://en.wikipedia.org/wiki/Out-of-band_management) console, relying on the currently installed OS to install a replacement can be tricky, as you already experience – HBruijn May 27 '19 at 07:32
  • Yeah, but that does actually require a connection to the management port, which I don't have unfortunately. Anyways, I looked again at Michael's answer and I edited the grub.conf to this: title Install CentOS 7 root (hd0,0) kernel /boot/vmlinuz inst.ks=http://server-ip/openvz7/example.ks initrd /boot/initrd.img – HudsonHawk May 27 '19 at 07:43
  • A proper out of band management system would allow you to use installation media such as ISO images which are on your own computer. Inserting physical media into the server is not necessary. – Michael Hampton May 27 '19 at 16:46