Follow Up:

So after doing more research, etc, I've discovered the following:

- Because the install failed the V-Bridge (virbr0 - and the 
Local VM (HostedEngineLocal - still exist on the Host.
- SSHing into the Local VM I discovered that I could ping out to an ip address, 
but not to a hostname (I used, as the initial error in 
this run was complaining that the Local VM couldn't download the 
centos-ceph-pacific meta-data).
- The Local VM's /etc/resolv.conf file has a single entry pointing to the 
- I did a "dig" to and that command returned the correct 
ip addresses, etc, and used the V-Bridge to do it.

HOWEVER, with some further experimentation (eg doing 2 pings to the same host 
in a row, as well as other experiments) the underlying issue *seems* to be that 
the dns requests are timing out - once the Local VM has resolved the 
hostname<=>ip address mapping everything *seems* to run OK.

So, I'm wondering if any of the following solutions are viable (of course, I 
could be completely off-track):

- Is there some way of adding the DNS servers that the deployment script asks 
about to the Local VM's /etc/resolv.conf file (ie bypass the V-Bridge when 
doing a DNS lookup to make the response time faster)?
- There is a `timeout` setting (default: 30 sec) in dnf config file. Is there 
some way of changing this to a larger value on the Local VM.
- Use the `--ansible-extra-vars=he_pause_host=true` (or similar) flag to pause 
the install, go into the Local VM and make one or both of the above changes.

If *anyone* has any ideas I'd love to hear them.


Users mailing list --
To unsubscribe send an email to
Privacy Statement:
oVirt Code of Conduct:
List Archives:

Reply via email to