You are here

Opsview Atom repository key not working

9 posts / 0 new
Last post
AGS Admins
agsadmins's picture
Opsview Atom repository key not working

Hello,

I have dropped an Opsview Virtual Appliance into Hyper-V and everything is running great. I have added my API key to my sources and I am able to run sudo apt-get update without any erorrs. However, when I run apt-get upgrade, I am not able to download any/many Opsview packages, as the Opsview repository server is responding with HttpError401. Could it be failing because apt-get is trying to install opsview-servicedesk-connector, which may not be a part of Opsview Atom?

Also note that:

  1. This forum does not actually support the <code>code</code> tags that it claims to support.
  2. I am unable to post the apt-get output as your spam filter thinks it is spam (which, obviously, it is not).
  3. If Opsview worked anything like this awful forum software, neither of them would ever again find their way inside my organization.

 

Duncan Ferguson
dferguson's picture
There was an issue with your

There was an issue with your account on our backend systems - please try the upgrade again.

Regarding your forum comments,

  1. You should be able to use the 'Styles => Computer Code' dropdown while using the WYSIWYG editor
  2. We have amended our SPAM filters so posting the apt-get output should no longer be an issue
  3. We are always looking at making improvements to the services we provide, so thank you for your comments.

Please let me know if any further issues

  Duncs

AGS Admins
agsadmins's picture
Perfect, I can upgrade

Perfect, I can upgrade without any apt-get issues now. Thanks Duncs! There is one issue that would be fairly critical though: After the upgrade, my host groups are not showing up on the default screen. That is, on the <servername>/monitoring/#! page, it simply says "No Hosts Groups available". If I go into the host group settings I can see all of my exisitng host groups and their servers, which is good, but they're not showing up on that screen and I'd really like to get them back. Any advice?

AGS Admins
agsadmins's picture
I can't see my comment here

I can't see my comment here after I refreshed the page, so I'll go ahead and post it agian:

Perfect, I can upgrade without any apt-get issues now. Thanks Duncs! There is one issue that would be fairly critical though: After the upgrade, my host groups are not showing up on the default screen. That is, on the <servername>/monitoring/#! page, it simply says "No Hosts Groups available". If I go into the host group settings I can see all of my exisitng host groups and their servers, which is good, but they're not showing up on that screen and I'd really like to get them back. Any advice?

Duncan Ferguson
dferguson's picture
Re your comments not

Re your comments not appearing - this is because all comments on threads wait on approval as an anti-spam measure.  I am not sure this is requried at the moment, so I'll discuss with the site admin.

Are you still missing the hosts from the monitoring pages?  Was this immediately after the upgrade and reload?  Are all daemons up and running okay (as nagios run 'opsview_watchdog summary')

  Duncs

AGS Admins
agsadmins's picture
I realized that it was

I realized that it was waiting for approval just after posting the second one. My bad.

And yes, just after the upgrade finished and I rebooted it was like that. I ran apt-get upgrade and everything installed without error. I rechecked the Opsview Updates warning to see if I was running the new version, but it still said I needed to upgrade, so I restarted the server to see if it would start the new version. When the server came back up it was not showing any information about my hosts. I cannot see my servers in the host groups view, the hashtag view doesn't find anything, graphs don't find anything to graph -- it's like none of the Opsview interface is seeing my configuration. However, when I go into the settings, all of my hosts and service checks are still there. I can even test the service checks and they succeed, so it's like the issue is that the frontend can't use the backend. 

I am not sure if all of the daemons are running. I tried to check the logs to see if there was anything obvious but couldn't see anything that stood out. Is there a list of what should be running somewhere? I can spin up the VM and check what stopped. But for now I reverted the VM to a previous checkpoint and only lost the last two weeks of performance data and alarm history, 

Duncan Ferguson
dferguson's picture
Can you bring up your VM and

Can you bring up your VM and check to ensure all packages have been fully upgraded but using the command on https://knowledge.opsview.com/articles/opsview-monitor-520/3523-upgradin... ?

Can you also post the output of running 'opsview_watchdog summary' as the nagios user a few minutes after bringing it up?  For reference, on my test system I see

$ opsview_watchdog summary
+----------------------------------------+---------------+-------------------+
| Service                                | Status        | Monitoring Status |
+----------------------------------------+---------------+-------------------+
| Process 'opsview-web'                  | Running       | Monitored         |
| Process 'opsview-timeseriesrrdupdates' | Running       | Monitored         |
| Process 'opsview-timeseriesrrdqueries' | Running       | Monitored         |
| Process 'opsview-timeseriesenqueuer'   | Running       | Monitored         |
| Process 'opsview-timeseries'           | Running       | Monitored         |
| Filesystem 'rootfs'                    | Accessible    | Monitored         |
| Filesystem 'varfs'                     | Accessible    | Monitored         |
| Filesystem 'optfs'                     | Accessible    | Monitored         |
| Process 'opsviewmd'                    | Running       | Monitored         |
| Process 'import_ndologsd'              | Running       | Monitored         |
| Process 'import_perfdatarrd'           | Running       | Monitored         |
| Process 'opsviewadmd'                  | Running       | Monitored         |
| Process 'nsca'                         | Running       | Monitored         |
| Process 'nrd'                          | Running       | Monitored         |
| Process 'opsviewnfd'                   | Not monitored | Not monitored     |
| Process 'nagios'                       | Running       | Monitored         |
| Process 'opsviewd'                     | Running       | Monitored         |
| Process 'opsviewhd'                    | Running       | Monitored         |
| Process 'opsview-agent'                | Running       | Monitored         |
| Process 'import_ndoconfigend'          | Running       | Monitored         |
| System 'djf-opsview'                   | Running       | Monitored         |
+----------------------------------------+---------------+-------------------+

  Duncs

AGS Admins
agsadmins's picture
It worked!

I ran the update command on the page you linked to -- "sudo apt-get install $(dpkg -l | awk '/opsview/ {print $2}')" -- and everything seemed to work as expected. After the upgrade I rebooted and it had my hosts how I left them yesterday. The only issue now is that SNMP was turned off on some hosts and I can't get SNMP to work again for my Linux hosts, though Windows hosts are responding to SNMP as expected.

It's strange that apt-get upgrade is not able to update all of the required packages. I ran the command above and the below apt output is what it found to install, meaning that not only did apt-get upgrade cause Opsview to stop functioning in a strange way, it failed to install most of what needed to be upgraded to begin with. 

Output of sudo apt-get install $(dpkg -l | awk '/opsview/ {print $2}') after apt-get upgrade had already been run:

The following packages were automatically installed and are no longer required:
  libapache2-mod-php5 libgd2-noxpm-dev lsof php5-cli php5-common php5-gd
  php5-json php5-mysql php5-readline
Use 'apt-get autoremove' to remove them.
The following extra packages will be installed:
  libossp-uuid16 opsview-external opsview-timeseries
  opsview-timeseries-enqueuer opsview-timeseries-lib opsview-timeseries-rrd
  uuid
The following NEW packages will be installed
  libossp-uuid16 opsview-external opsview-timeseries
  opsview-timeseries-enqueuer opsview-timeseries-lib opsview-timeseries-rrd
  uuid
The following packages will be upgraded:
  opsview opsview-core opsview-setup opsview-web

Given that apt-get upgrade is the standard software update mechanism in Debian systems, I would suggest that you make it abundantly clear that doing so can completely break Opsview whenever someone runs apt-get upgrade, and then suggest the above code to upgrade correctly. 

Here is the current output of opsview_watchdog summary on my system:

(can't post it, it's "spam"...)

Duncan Ferguson
dferguson's picture
I will get the spam filters

I will get the spam filters looked into again - thanks for letting me know about that

As for using apt-get upgrade, I would suggest always using apt-get dist-upgrade in the future (I always use this without any problems) as there is significant difference between the two; 'upgrade' does not install extra packages or remove redundant ones (it only upgrades what is installed) whereas 'dist-upgrade' does.

Regarding the Linux SNMP checks; which checks are these?  Does 'Test Check' work as expected on the Host edit SNMP tab?  Do you have any extra settings such as 'Max Message Size' set?

  Duncs