Experience on armv7 install

arm

(Robert Moskowitz) #1

I am posting to development as this is on Centos-7 arm.

I did a centos-gnome image install on a sata drive for a Cubietruck board. I set up a user account and installed tiger-vncserver and got gnome working over vnc with the system in multi-user. I set up firewalld for port 5901 and 5902.

I then followed the instructions to start the install.

I answered NO and went back and set the hostname and timezone. Later I learned the install would do this for me. The install reported:

This is your current system setup:

  Size SD-Card: 7.97GB

Free space SD-Card: 7.97GB
Size RootFS: B
Available RootFS: B
Max CPU Frequenty: 960 MHz

Static hostname: klovia.htt-consult.com
Icon name: computer
Machine ID: b68e584ad1e9435289d9bf108bbc515e
Boot ID: 9ac6c8818ae64aac8b883e0c777f97c3
Operating System: CentOS Linux 7 (AltArch)
CPE OS Name: cpe:/o:centos:centos:7
Kernel: Linux 4.14.71-201.el7.armv7hl
Architecture: arm

     Time zone: America/Detroit (EDT, -0400)
 System Locale: LANG=en_US.UTF-8

where did it get the size? I have the OS on a 320GB HD…

at the beginning of the install I caught:

Package deltarpm-3.6-3.el7.armv7hl already installed and latest version
Nothing to do

NO DELTA RPMS on arm, particularly IF the install is on SD card. In my case it would not matter too much, but it seems that no delta rpms were downloaded.

Now a couple notes about the install.

It deleted my /home/me account. I lost everything I had there. It left the /etc/passwd and group alone. That will cause problems later…

It removed firewalld for shorewall (I haven’t used shorewall in over a decade!) and lost the rule for ports 5901 and 2 for vnc. But the service is still running; which it won’t do after a boot and ~/.vnc/xstartup is gone.

The last message from the install was:

You can access the Web interface at:

https://klovia.htt-consult.com:980 (or https://192.168.192.14:980)

Login: root
Password: <your_root_password>

but I could not connect. From this notebook, I would have to add a firewall rule for port 980 (different VLANs). Fortunately I had a F29-beta arm Cubie on the same VLAN. It could not connect so I checked the addressing:

ip a

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
link/ether 02:c1:08:81:e2:d7 brd ff:ff:ff:ff:ff:ff
inet 192.168.192.14/24 brd 192.168.192.255 scope global noprefixroute dynamic eth0
valid_lft 86148sec preferred_lft 86148sec
3: wlan0: <BROADCAST,MULTICAST> mtu 1500 qdisc pfifo_fast state DOWN group default qlen 1000
link/ether 98:3b:16:1e:e1:e0 brd ff:ff:ff:ff:ff:ff
[root@localhost ~]# ping 192.168.192.12 # this is the system trying to connect…
connect: Network is unreachable
[root@localhost ~]# route -n
Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref Use Iface
169.254.0.0 0.0.0.0 255.255.0.0 U 1002 0 0 eth0

what happened to the routing table???

So:

ifdown eth0;ifup eth0

Determining IP information for eth0…[ 2198.790032] RTL8211E Gigabit Ethernet stmmac-0:00: attached PHY driver [RTL8211E Gigabit Ethernet] (mii_bus:phy_addr=stmmac-0:00, irq=POLL)
[ 2198.816200] sun7i-dwmac 1c50000.ethernet eth0: RX IPC Checksum Offload disabled
[ 2198.823877] sun7i-dwmac 1c50000.ethernet eth0: No MAC Management Counters available
[ 2198.831821] sun7i-dwmac 1c50000.ethernet eth0: PTP not supported by HW
[ 2198.839514] IPv6: ADDRCONF(NETDEV_UP): eth0: link is not ready
[ 2200.928849] sun7i-dwmac 1c50000.ethernet eth0: Link is Up - 100Mbps/Full - flow control off
[ 2200.965803] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
done.
[root@localhost network-scripts]# route -n
Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref Use Iface
0.0.0.0 192.168.192.1 0.0.0.0 UG 0 0 0 eth0
169.254.0.0 0.0.0.0 255.255.0.0 U 1002 0 0 eth0
192.168.192.0 0.0.0.0 255.255.255.0 U 0 0 0 eth0

And I could connect to the web interface and continue the install.

I have to logoff now until probably Tuesday evening when I will attempt to report what happened during the setup and app installs.

More to come…


(Mark Verlinde) #2

I did not know centos-arm images with a desktop are distributed…
Aim of Nethserver is a web-based user interface without (X11) gui.

It reports the size of the sd card. Asuming you have just u-boot on your SD and the root fs lives on the HDD this is not very relevant. It is in place so users can judge if they want to “grow” their root-fs before engaging the installation. We should do a better job here…

Have to check this, I cannot recall we are trying to touch /home

It should not be necessary to do this manually. , at fisrt boot it should try to get an IP from a DHCP server.
Still thinking / wondering what went wrong here. :thinking:


(Robert Moskowitz) #3

boy, did the email response come up wrong! Here it is properly formatted.

On 9/29/18 1:54 AM, Mark Verlinde wrote:

\ 45x45 mark_nl Mark Verlinde Arm Team Coordinator
September 29

rgmhtt:

I did a centos-gnome image install on a sata drive for a Cubietruck board

I did not know centos-arm images with a desktop are distributed…
Aim of Nethserver is a web-based user interface without (X11) gui.

With the latest image: http://mirror.centos.org/altarch/7/isos/armhfp/

You have minimal, GNOME and KDE. And I want Xfce, but not all the needed modules are there.

I know that you go for the web based interface. I have always found having alternative access allows me to easily do other things. And I like more than just SSH access. By using VNC, there is very little X11 overhead on the server.

rgmhtt:

where did it get the size? I have the OS on a 320GB HD…

It reports the size of the sd card. Asuming you have just u-boot on your SD and the root fs lives on the HDD this is not very relevant. It is in place so users can judge if they want to “grow” their root-fs before engaging the installation. We should do a better job here…

You should look at the /dev device where / resides.

rgmhtt:

It deleted my /home/me account.

Have to check this, I cannot recall we are trying to touch /home

I was very surprised. Like what happened here? Again, if you start with a Centos image, you may well have previously set up things before the Nethserver install. And with the GNOME and KDE images, you pretty much have to set up a user account. Deleting /home is bad. and then when you go to add a user, if the /home/user already exists, just accept it.

rgmhtt:

ifdown eth0;ifup eth0

It should not be necessary to do this manually. , at fisrt boot it should try to get an IP from a DHCP server.
Still thinking / wondering what went wrong here. :thinking:

Hmm, first boot? that was the Centos image setup. THEN you install Nethserver, already past first boot. Should not need a 2nd boot. Looks like you built an ifcfg-eth0 file and that may have messed things up.

I will check my notes for other events of the install.


(Mark Verlinde) #4

I am not sure if it is even possible a DE and Nethserver can coexist. Doubt that (as an example) : we disable network manager to have full control over network settings.


(Robert Moskowitz) #5

ClearOS also disables NM, and they ship with GNome. This is not an issue. Do you still use the nmcli for setting up your ifcfg files and such? You really should if you don’t.

Do you provide an edit option for crontab? Yes, I can do that in SSH, I just like how much easier it is in a terminal window on the server.

also I test out my rsync commands in a terminal window.,

Finally, the biggest is using the GUI file manager to figure out what my wife has done with her home directory!


(Robert Moskowitz) #6

And with a GUI console, you can always run your browse of choice and connect to localhost:980 for the web admin. Again ClearOS provides this and even provides a desktop icon that directly connects you. I also remember long ago with AMAHI doing this.


(Robert Moskowitz) #7

I succeeded in installing the gnome desktop and then accessing it via vncserver. Both for a user I created and for root. Took me a little bit to figure out shorewall to open ports 5901:5902.

Too much memory and CPU (IMHO) for arm. Plus I have pretty much forgotten how to get around gnome.

So instead XFCE. There are a couple problems. First for some reason the “Xfce Desktop” group that is available in x86_64 Centos7 is not in Centos-arm.

Then back in August, I tried hacking at doing this manually by getting help on what is in the XFCE ande X11 desktops and doing those manually. Not all the rpms are available. Worst, back in August I got install errors. You can see my posts on this on the Centos arm list, and, for now, will not repost them here.

Hopefully someone is interested in helping me work on this. I can fall back to webmin to get some added server maint tools, but would prefer a DE.