Lab-8.1. Packstack - OpenStack Dashboard is not Available after Successful Install

vipinsagar

Lab-8.1.
Packstack install completed successfully:

And I don't see a Public-IP

I can't logon to OpenStack BUI and OpenStack Dasboard tab shows the apache welcome page. I can source the keystone and run CLI

How can I access the OpenStack BUI,

vipinsagar

Additionally:
One of the prep is yum update and it install a new Kernel and if you rebooted the server to get the new kernel to boot up, you will loose the instance and all the update. Possibly clear the cache or open the lab in incognito mode.

serewicz

The 172.x IP address is acting as your public IP address. If you read the paragraphs and graphics at the start of each chapter you'll find they explain the use of the tab inside the Katacoda page, with the title OpenStack Dashboard as well as the URL under /opt/host to gain access to the BUI.

I will test the issue after update. This is not the expected happening, I'll post after I figure out the issue.

Regards,

serewicz

Could you tell me which step had the reboot command that caused you issues?

vipinsagar

There is no reboot step in the Lab, but there was new kernel package part of the yum update. And I was thinking to run reboot or systemctl reboot to boot into the new kernel. After the reboot: I see the below message

Your Interactive Bash Terminal.*
>

ssh: connect to host host01 port 22: No route to host

>
>

The environment has expired.

>

Please refresh to get a new environment.

>
>

We encountered a problem with the connection.
This may have been due to the internet connection dropping.

>

Please refresh and try again.

>
>
I did refresh but keep getting the same error message, even after couple of hours thinking boot up may take some time.

serewicz

Yes. There isn't a reboot in the lab steps to avoid this issue. After adding the new repo the update will update the other files which should allow the rest of the lab to run smoothly. Continue without the reboot.

Regards,

vipinsagar

@serewicz said:
The 172.x IP address is acting as your public IP address. If you read the paragraphs and graphics at the start of each chapter you'll find they explain the use of the tab inside the Katacoda page, with the title OpenStack Dashboard as well as the URL under /opt/host to gain access to the BUI.

I will test the issue after update. This is not the expected happening, I'll post after I figure out the issue.

Regards,

I did access by following as per the lab doc: Use the OpenStack Dashboard tab via the Katacoda page to access the Horizon BUI.

Before the install, it will ask to specify the port number, the NGINX page.
After successful RDO install, the dashboard BUI access shows the apache welcome page directly.

I am going to run/install the packstack again. Will see again the dashboard and /opt/host

serewicz

If you see the Apache welcome page this typically means the packstack has not finished, or there was an error. It looks like the install worked from the output. It can take a minute or two for all the OpenStack services to finish configuration.

The output at the end of the log file is listing the 172.x.y.x.z IP address which Packstack is using to present Horizon. As we are connecting to the nodes remotely via the Katacoda page we will need to use a different URL.

Regards,

vipinsagar

OK Install finished.

There was not file/dir host under /opt

Regards,

serewicz

I cannot seem to get lab 8 to come up at the moment myself, it says there is a network issue and the connection drops. I have reached out to Katacoda to see if they can look at the issue. I'm not sure when they will get back to me, but will post here as soon as I can.

Until then I was able to get lab 10 to come up, and it looks to be working properly. Once lab 8 is fixed you could return.

Regards,

serewicz

Hello again,
Yes, I found your same issue. The /opt/host file was not found. I will work to the lab support folks and update the lab steps. Until then please use the URL found after you click on the OpenStack Dashboard link and add a ServerAlias with that path. Then restart the httpd and memcache services. This is what worked for me, but your URL will be slightly different:

[root@a644c4b092c1 ~]# vim /etc/httpd/conf.d/15-horizon_vhost.conf
[root@a644c4b092c1 ~]# grep ServerAlias /etc/httpd/conf.d/15-horizon_vhost.conf
ServerAlias 172.17.0.14
ServerAlias a644c4b092c1
ServerAlias 35.201.124.219
ServerAlias 2886795278-80-ollie03.openstack-environments.katacoda.com ## <----Add URL from other tab here
ServerAlias localhost
[root@a644c4b092c1 ~]# systemctl restart httpd
[root@a644c4b092c1 ~]# systemctl restart memcached

Thank you for bringing this to our attention.

Regards,

vipinsagar

No Problem. Happy to help. Thanks for addressing the issue and taking the action to fix.
Lab-10 pretty much covers for Lab-8.
I will try adding the dashboard URL as one of the ServerAlias next time.
Best,

vipinsagar

SOLVED

I see the Lab-8 is updated with ServerAlias host and /opt/host and I can confirm the issue is fixed with new lab version for Lab-8. I am able to launch BUI. Thanks again adressing the issue.

serewicz

Glad it worked for you.

Regards,

eddelsi

Hi all,

I am executing now also Lab 8.1 and getting only to the apache testing page. I added the ServerAlias in the httpd.conf
file as described in the lab and restarted the respective services. Could you please help?

Thanks!

serewicz

Hello,

When the packstack command finished there should have been output indicating successful completion of the script, including how long it took for it to run. Did you get that output?

When you select the tab for "OpenStack Dashboard" what do you see?

At the end of the log file do you see any errors?

Which IP address are you adding as the ServerAlias, the external IP you would use to SSH to the instance?

Regards,

serewicz

Hello,

In the most recent screen shot it appears that you put in a 172.17.0.58 IP address. This address is not one that routes across the internet, and would be internal to the lab environment. Which is why the URL found under the browser tab is that long path ending in katacoda.com.

As this is a fresh installation you'll note that the ScriptAlias entries may not yet be populated. Please check that the URL found under the OpenStack Dashboard tab is listed as a ScriptAlias entry and that httpd and memcached have been restarted.

vim /etc/httpd/conf.d/15-horizon_vhost.conf
systemctl restart httpd ; systemctl restart memcached

When you open a different lab do you find the dashboard works?

Regards,

MikeRiggs

Hello all,

I am running into this same issue with this lab. Packstack completes successfully with no errors and I have edited/verified the files as described in this thread as well as restarted httpd and memcached. Any thoughts on what else to look for?

Thanks for any help.

serewicz

Hello,

To understand you are saying you see the test page and not the openstack login page? Is this true when you open the "OpenStack Dashboard" tab?

Just want to be sure I'm troubleshooting the correct issue.

Regards,

serewicz

Hello,

I have looked into the issue. It seems the network configuration of the lab nodes has changed since the labs were written. I have reached out to the maintainer of the lab to inquire if this is an ongoing change, or is an unintended update.

Until this is figured out, please look in the /opt/host file, and use the node name listed there as the ServerAlias to use. This is mentioned in the intro to the lab, but the name should match an external IP address found with ip a. The node no longer shows an external IP, but an internal one which causes the ServerAlias to not match the request and you are finding the default, test web server page. On the instance I just tried it went like this, your name may be different:

[root@rdo-cc ~]# cat /opt/host
http://2886795298-80-ollie03.openstack-environments.katacoda.com/
[root@rdo-cc ~]# vim /etc/httpd/conf.d/15-horizon_vhost.conf
....
 ServerAlias 2886795298-80-ollie03.openstack-environments.katacoda.com
.....

[root@rdo-cc ~]# systemctl restart httpd ; systemctl restart memcached

MikeRiggs

I appreciate the thoughts. I have tried the /opt/host file fix and the ServerAlias fix mentioned in this thread and I still get the test page rather than the OpenStack login page. I'll try it again tonight just to make sure.

serewicz

Hello,

I will be updating the labs to match the new network configuration. I tested the steps where you find the external URL and create a ServerAlias entry for it. One note, be sure not to use the https:;// portion, only the name. I had done a simple copy/paste the first time and got the same test page. The alias should be just the name, no other notation.

Regards,

MikeRiggs

Thanks for the update. I have not had a chance to test this yet, but I know that when I tried the ServerAlias entry in the past I put the https:// part in there. I'll try it without when I have time to do so.