Newer fedora is working on dropping scp (for lots of good reasons)
internally, but using a sftp wrapper. ie, you call scp and internally
ssh uses sftp. This won't work unless the server has sftp enabled.
I (and I suspect other admins) scp things to batcave from time to time
so it would be nice to have this on.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
This is a trial run to see if it's viable for these hosts. I
probably got something wrong. Let's see what explodes.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
BLOCKERBUGS_URL needs to be defined in order to show correct links in the
staging environment. While fixing that, realized that most of the if-else
clauses were unnecessary, because there are already group vars for that. So
simplified the config file by using the vars. KOJI_URL is dropped because it's
no longer needed.
Related: https://pagure.io/fedora-qa/blockerbugs/issue/192
The box is repeatedly getting stuck and requiring a reboot. It
didn't do so without the workers running, so possibly it's just
overloaded. Let's try it with 20 workers.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
Also add a ssl connection cache.
These changes are live on proxy01/10 and seem to have made them stable
again. Will look at pushing to the rest tomorrow.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
* Only support one Fedora release (34) on staging retrace.
* Update list of removed versions on staging.
* Support CentOS 7 and 8 on production instance.
* Update repository URLs, remove dead links.
The cert here is gotten from centos.org folks and checked into private.
At some point we will get a longer term cert here so there's less manual
churn. Otherwise we just make a new mirrors.stg.centos.org and proxy it
to mirrorlists just like mirrors.stg.fedoraproject.org.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
By default nagios assumes it can connect to everything in inventory via
nrpe and monitor things like swap and number of emails in the postfix
queue. For ocp hosts running CoreOS we don't want to have nagios
monitor any of that, we only want it to monitor ping (is the host up).
This change is only in vars here, but it needs a noc run to activate, so
it will need a freeze break to run the noc playbook (as noc is frozen).
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
This host is an old phx2 host thats been sitting there on this ip, but
we haven't moved any vm's to it (thus it shouldn't be affected by the
freeze). However, we should get in ansible so it's network is managed
and so we remember to reboot it for updates, etc.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
