[Openstack] [Fuel] Cannot access Fuel 10.0 UI because of unavailable Nailgun server

Vincent Jahjah vincent_jahjah at hotmail.com
Thu Feb 16 14:23:57 UTC 2017


Hello, thanks for the intel.


Sadly, I have tried the ISO installation way, and for some reason, any Fuel ISO installation fails during the earliest stages. Our IBM (IBM x3650 M3) server seems to be the issue (Something about Dracut init having timedout), as on our HP server the installation seems to work (although the guys at IT refuse to let me use it!)


Interestingly, installing on the HP server had a different kind of error, which appears to be either a bug in the USB-stick burning, or in the Fuel ISO. The error given was "Cannot fetch kickstart file from [USB-stick name]/ks.cfg"

As it happens, the path name was wrong in the syslinux.cfg file (if I'm not mistaken; it was a cfg file with paths to ks.cfg). Apparently, the name of the USB-stick did not match, so ks.cfg could not be found. Manually changing the path name solved the issue.


But none of that is enough to get my fuel to work...


Vincent

________________________________
De : Amit Kumar <ebiibe82 at gmail.com>
Envoyé : 16 février 2017 06:47:26
À : Vincent Jahjah
Cc : openstack at lists.openstack.org; Evgeniy L
Objet : Re: [Openstack] [Fuel] Cannot access Fuel 10.0 UI because of unavailable Nailgun server

Hi Vincent,

Earlier we were doing installation using fuel-community-10.0 from following link:
https://www.fuel-infra.org/#fuelget

For installation, we were following the steps ("If you choose installation using pre-built ISO image") as mentioned in:
https://docs.openstack.org/developer/fuel-docs/userdocs/fuel-install-guide/install/install_install_fuel_master_node.html

To our surprise, we were not getting fuelmenu for the configurations and when installation was getting done. It was not complete, no FUEL UI etc, no network connectivity to machine whereas we had provided one of the NIC on external NW. Problem was that we were not getting fuelmenu during installation process to configure network setting etc. for smooth installation.

Then we followed the steps ("If you choose installation using RPM packages then follow the procedure described here") mentioned in the same link mentioned above. On these steps, we faced the same cobbler sync failure messages. To our doubt, it was definitely seems to be some issue with the latest fuel RPM that we got as per the above mentioned steps.

We again switched back to installation from ISO image now knowing the script for bootstrap. We let it went as it was going. Same problems were there as faced earlier. But once installation was done, we ran sudo /usr/sbin/bootstrap_admin_node.sh again. fuelmenu was there during installation, we made sure network connectivity is up. Follow this: https://docs.openstack.org/developer/fuel-docs/userdocs/fuel-install-guide/install/install_change_network_interface.html. After this installation of FUEL master node was successful.

Hope it helps.

Regards,
Amit


On Thu, Feb 16, 2017 at 9:09 AM, Vincent Jahjah <vincent_jahjah at hotmail.com<mailto:vincent_jahjah at hotmail.com>> wrote:

Hello,


I have not. The techies are currently considering trying on an entirely different server and hoping for the setup to work. They told me my setup is sound, though, so I really have not the least idea what could be wrong. Have you tried reproducing the steps I have listed in my previous mail?

________________________________
De : Amit Kumar <ebiibe82 at gmail.com<mailto:ebiibe82 at gmail.com>>
Envoyé : 15 février 2017 20:47:32
À : Vincent Jahjah
Cc : openstack at lists.openstack.org<mailto:openstack at lists.openstack.org>; Evgeniy L

Objet : Re: [Openstack] [Fuel] Cannot access Fuel 10.0 UI because of unavailable Nailgun server

Hi Vincent,

Have you found solution to your problem? I am facing exactly same problem now at my setup. Any pointers?

Thanks.

Regards,
Amit

On Feb 14, 2017 8:20 PM, "Vincent Jahjah" <vincent_jahjah at hotmail.com<mailto:vincent_jahjah at hotmail.com>> wrote:

Fortunately, I have been documenting my step-by-step way of installing Cent OS.


1. Download a CentOS ISO (in our case, CentOS-7-x86_64-Everything-1611.iso)
2. Burn the ISO on a USB stick (in our case, using rufus-2.12.exe)
3. Set the server to boot on the USB stick and proceed to the installation
    (In our case, on an IBM server)
4. Modify the following files to setup the network of CENTOS:

vi /etc/sysconfig/network-scripts/ifcfg-AAAX
    (In our case, AAAX = eno1)

DEVICE=eno1
BOOTPROTO=static
BROADCAST=10.12.99.255
HWADDR=E4:1F:13:68:3F:28
IPADDR=10.12.99.34
NETMASK=255.255.255.0
NETWORK=10.12.99.0
ONBOOT=yes

vi /etc/sysconfig/network

NETWORKING=yes
NETWORKING_IPV6=no # this line does not appear to matter
HOSTNAME=lab.uni.myuniversity.ca<http://lab.uni.myuniversity.ca>
GATEWAY=10.12.99.2

vi /etc/resolv.conf

nameserver 8.8.8.8

5. Type the following command to restart the network

service network restart

6. ping 8.8.8.8 should work (it normally doesn't work until all previous steps are completed)

Now, follow the instructions to install FUEL on top of CENTOS 7, as described here:
docs.openstack.org/developer/fuel-docs/userdocs/fuel-install-guide/install/install_install_fuel_master_node.html<http://docs.openstack.org/developer/fuel-docs/userdocs/fuel-install-guide/install/install_install_fuel_master_node.html>
These instructions are synthesized thus:

7. sudo yum install -y PyYAML net-tools
8. sudo rpm -i http://packages.fuel-infra.org/repositories/centos/master-centos7/os/x86_64/Packages/fuel-release-10.0.0-1.mos6377.noarch.rpm

(this step may result in a warning (NOKEY); we ignore it)

9. sudo yum install -y fuel-setup

Finally, start the fuel bootstrap installation
(this step should take ~5 minutes):

10. sudo /usr/sbin/bootstrap_admin_node.sh

Once you enter the fuel configuration menu...

11. Change the password, which is currently admin ...

12. Verify that the specifications in "Network Setup" corresponds to the network specs you described earlier.

13. The DHCP pool should look something like 'eno1' : Start at 10.12.99.3, end at 10.12.99.254, Gateway 10.12.99.34

In our case, no modifications are necessary.

14. IF you have a running DHCP service, it must be stopped or it will conflict with the fuel one.
In our case, 10.12.99.2 was a DHCP service. Putty can be used to identify and delete the service:

sh run | inc dhcp
    -> The output will allow you to find the relevant dhcp pools. Note the name.
no ip dhcp [name of the dhcp pool]
do wr

15. Proceed to install fuel 10.0 (Quit Setup -> Save and Quit)

-> Everything functions normally up to this point.
-> Crashes after 10 minutes, with the previous Cobbler sync failure messages.


________________________________
De : Evgeniy L <eli at mirantis.com<mailto:eli at mirantis.com>>
Envoyé : 13 février 2017 14:09:59
À : Vincent Jahjah
Cc : openstack at lists.openstack.org<mailto:openstack at lists.openstack.org>
Objet : Re: [Openstack] [Fuel] Cannot access Fuel 10.0 UI because of unavailable Nailgun server

I have several additional questions:
1. What version of Fuel do you use?
2. What instruction do you fallow to install Fuel?

>> reinstalling from a "basic" CentOS 7 installation
3.  Could you please clarify how exactly you do installation?

Thanks,

On Mon, Feb 13, 2017 at 9:33 AM, Vincent Jahjah <vincent_jahjah at hotmail.com<mailto:vincent_jahjah at hotmail.com>> wrote:

So, I tried reinstalling from a "basic" CentOS 7 installation.

I observed similar strange behavior from the previous time I tried the installation, which means it is reproducible and that I don't can't move on.


I don't know how to export the log file from my server, and considering its size, I doubt it's much of a loss if I send only screenshots of the most suspicious parts. The last one specifically is the Cobbler failure, which has a stacktrace that doesn't seem to appear in the log file; it only appears during the error (and I can't scroll up D:).


Reading the error, I checked if httpd was running, but "status" indicates it's active. I'm not sure what is meant by "SELinux could be in the way".


The other screenshots are in chronological order, the first one occurring shortly after the GRUB setup of fuel; I understand it has to do with finding apt-get-like programs.


(I'm sorry if my feedback output is limited: the cycle for reproducing the error is of about an hour and I'm very limited by my remote-access technology in terms of copy-pasting things. If you really need additional feedback, I'll do my best to give you the full output)


Regards

________________________________
De : Evgeniy L <eli at mirantis.com<mailto:eli at mirantis.com>>
Envoyé : 10 février 2017 17:58:56

À : Vincent Jahjah
Cc : openstack at lists.openstack.org<mailto:openstack at lists.openstack.org>
Objet : Re: [Openstack] [Fuel] Cannot access Fuel 10.0 UI because of unavailable Nailgun server

The problem is for some reasons your Fuel Master installation was not finished successfully, there are logs which you can check to figure out what went wrong (see my previous message), also send a full output of `bootstrap_admin_node.sh` command execution. It may make sense to try to install from scratch, if you did a lot of manual changes and configurations, to ensure there are no overlapping problems.

Thanks,

On Fri, Feb 10, 2017 at 9:42 AM, Vincent Jahjah <vincent_jahjah at hotmail.com<mailto:vincent_jahjah at hotmail.com>> wrote:

Hello,

Thank you for your advice!


Curl did not work on the master node.

It seems that nginx was not installed (service nginx status did not detect nginx).

I used yum to install nginx, but doing this seems arbitrary and does not seem to do anything (was it not supposed to be there in the first place?)


The bootstrap_admin_node logs suggested I use "fuel-bootstrap build --activate" once I had fixed the network.

This command crashes on a Cobbler issue (first image).

The configuration file of Cobbler has a comment suggesting I restart the cobbler demon. (second image)

The command still fails after restarting the demon.

This is the point where I begin to suspect httpd (apache) is in cause.

When I restart httpd, I get a strange error explaining how there is a name duplication in the Cobbler configuration file inside the /etc/httpd folders. (two last images) Specifically:


> # Use separate process group for wsgi
> WSGISocketPrefix /var/run/wsgi
> WSGIScriptAlias /cobbler_web /usr/share/cobbler/web/cobbler.wsgi
> WSGIDaemonProcess cobbler_web display-name=%{GROUP}  /// ***This line***
> WSGIProcessGroup cobbler_web
> WSGIPassAuthorization On


At this point, I am rather clueless and am starting to wonder if I am not breaking the OS further by stabbing in the dark. I wonder if this is leading anywhere, or whether I should just restart the Fuel installation from the CentOS image...


Vincent

________________________________
De : Evgeniy L <eli at mirantis.com<mailto:eli at mirantis.com>>
Envoyé : 7 février 2017 13:07:32
À : Vincent Jahjah
Cc : openstack at lists.openstack.org<mailto:openstack at lists.openstack.org>
Objet : Re: [Openstack] [Fuel] Cannot access Fuel 10.0 UI because of unavailable Nailgun server

Hi,

Try to access UI using "curl" from your machine and from Fuel Master:
1. If it's not available from your machine but available from Fuel Master, make sure that your network is configured correctly use regular troubleshooting techniques for that ping/traceroute/tcpdump.
2. If it's not available from both, your machine and Fuel Master, verify that Nginx up and running and that it's configured correctly, there could be problems with installation see "/var/log/puppet/bootstrap_admin_node.log" for details.

Thanks,

On Sun, Feb 5, 2017 at 6:21 AM, Vincent Jahjah <vincent_jahjah at hotmail.com<mailto:vincent_jahjah at hotmail.com>> wrote:

Hello,


I have attempted installing Fuel on my IBM server for testing purposes.

I currently find myself in a similar situation as the one described in this Openstack ask post: https://ask.openstack.org/en/question/84510/fuel-ui-network-error/


I can ping to and from the server, and when I try to connect to the UI from my web browser, the page does not load. The browser tells me I might have too stringent security requirements, but this does not seem to be the issue (deactivating the firewall does not change anything).


I cannot restart the Nailgun server simply because the dockerctl commands are deprecated on my version, and I have not found anywhere the commands that are meant to replace it.

fuel-utils reveals no issues with Nailgun, but apparently something is off with Cobbler (that's probably not relevant though).


When I check for the availability of the Nailgun server, I am told "Can't connect to Nailgun server! Please modify "SERVER_ADDRESS" and "LISTEN_PORT""


My config.yaml file looks like this:


SERVER_ADDRESS:"10.12.99.34"
LISTEN_PORT:"8000"
SERVER_PORT:"8443"
KEYSTONE_USER:"................"
KEYSTONE_PASS:"................"
KEYSTONE_PORT:"5000"

The IP is the host's IP as well as that advertised on the fuel console launch; 8443 is the advertised port for login as well.

I have added some of the following rules to the host's firewall, more or less blindly following tutorials:

*filter
:INPUT DROP [0:0]
:FORWARD DROP [0:0]
:OUTPUT ACCEPT [243:14376]
:ext-filter-forward - [0:0]
:ext-filter-input - [0:0]
-A INPUT -i lo -m comment --comment "000 allow loopback" -j ACCEPT
-A INPUT -p tcp -m state --state NEW --dport 23 -j ACCEPT
-A INPUT -m state --state NEW -m tcp -p tcp --dport 22 -j ACCEPT
-A INPUT -m state --state NEW -m tcp -p tcp --dport 8000 -j ACCEPT
-A INPUT -s 10.12.99.0/24<http://10.12.99.0/24> -p tcp -m multiport --dports 22 -m comment --comment "010 ssh" -m state --state NEW -j ACCEPT

I do "service iptable restart" and I restart firewalld after modifying this file.
I restart the network before trying to access the fuel UI.
I do all of my modifications on my client machine, on PuTTy using Telnet.

I've been stuck on this issue and have been on the lookout for support, hence my joining this email list.
This is not the only issue I've run into, but that's the most pressing one right now.
I know that I am attempting to join the right site on the browser because the tab icon on Chrome changes to the Fuel icon.


Regards,

Vincent


fuel ui network error - Ask OpenStack: Q&A Site for ...<https://ask.openstack.org/en/question/84510/fuel-ui-network-error/>
ask.openstack.org<http://ask.openstack.org>
Hi All I am trying Mirantis Fuel 7 or OpenStack deployment. Environment: Fuel Master - 2 IP (10.20.0.2, 192.168.x.y, Gateway - 192.168.x.z) Fuel slaves - 2 Using ...



_______________________________________________
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to     : openstack at lists.openstack.org<mailto:openstack at lists.openstack.org>
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack





_______________________________________________
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to     : openstack at lists.openstack.org<mailto:openstack at lists.openstack.org>
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20170216/e972b5c5/attachment.html>


More information about the Openstack mailing list