Over time out git repos get large and require some upkeep. Ideally
this should be done in the source-repositories element. We can revert
this code once we have landed that upstream.
Change-Id: I6b9ecd4448d97b5ab2bf3808ff36ff4e74be72e4
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
It needs to be tuned depending on number of cpus
and memory of the server.
Change-Id: Id00d32c1bc225db16fa45d990ff8227db4f68ff0
Depends-On: Id5a0e5af031175cc0a9d8b5a3bb97ae4adfbdf1e
We don't need the puppetmaster to send anything to puppetdb
since we don't use the puppetmaster any more -- reports are
injected into puppetdb with ansible now. Actual puppetdb
connections from puppet only happened with 'puppet apply' running
on the puppetmaster, making it undesirably special. Disable it.
Change-Id: I7df300d71ccfeb3698f5ccb97a73ba37ab4c57e0
Stop hardcoding the network settings on the modules
and instead of that, use parameters and rely on public
hiera.
Change-Id: I3cccfc03609bf90503b3fe27d3e717f89f6ec654
Allow nodepool to launch 3 process, instead of 1. This is to overcome
capacity issues with a single process.
Change-Id: I6a33359d8b488be6a44b09078401de02d68ef79e
Depends-On: Ibe41ce3c488da9cd68d8833b61fd42682f0b4e73
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
We need to replace cacti as it is struggling with memory issues.
Created cacti01 machine with 4GB, now adding the node declaration
to Puppet so it gets configured as a Cacti server.
Change-Id: I47a286079e44cd22686e99b8e4c229c693d509d1
The mqtt connections seem to be crashing the logstash daemons. So lets
revert this for the time being and when we fix the root cause we can
turn it back on.
This reverts commit eb72771d6f5ca6e21215b40118181ccb97859bec.
Change-Id: Ieaed43f5b94e0c5e2c98b0d627490d388e48261d
We're able to pretty reliably crash firehose with multiple websocket
connections at once. So to prevent us from DOS ourselves lets block off
the websocket ports for now. We can revert this when we have a remedy
in place.
Change-Id: I909ad4b160a152ae9b909a9e9a1e5d63afa39345
Now that logstash is working again lets turn on mqtt again.
This reverts commit c18db2ac089da0979d4bbecc47c138fde88bbd6d.
Change-Id: I6b395eb7786152c4ae4a1e0a6ba615340bfd237f
Logstash is busted and the last time it updated elasticsearch was when
we attempted to turn this plugin on. Whose bright idea was it to use
that anyway. This commit turns off the configuration for the mqtt output
plugin and reverts us to just using the elasticsearch output for now. We
can investigate adding this again at some point in the future.
Change-Id: I88d5e507c8b32a6969a48d6bf6e6881985898387
We have one bifrost machine to provision both infracloud flavors,
therefore the hiera key containing the ipmi passwords is the same
for all servers.
Change-Id: I9d69330f453af21a063f6585d17f91bc3925fc8e
This commit will enable the mqtt output plugin on logstash workers. So
now we'll be emitting logstash events to the firehose for anyone to
listen to.
Change-Id: I570a461ee13f5dfa5494554df1cc321fc6cdbf6c
Depends-On: I7255f6c256ab3e3ca33caf69b71bf4ffab02c1bb
The imap server on firehose.openstack.org doesn't actually need
external connectivity since the lpmqtt daemon is running on the same
host. This commit removes the imap ports from the firehose port list.
Change-Id: I5a9aaafe566dbd3e4e2e2f5790aec812db99f0ee
This is required to publish doc jobs to AFS with zuul.
Change-Id: Icfa8f83906f97e0ca98a240c0514289edacba0fb
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
This commit adds deploying a running lpmqtt to firehose nodes.
Change-Id: Ia9d659e282a2a992b8c1a7a48577f3e59793effa
Depends-On: I613330e2bff2e6fe1cacd7e53f3c189584978ea2
A requirement for running lpmqtt is to have an imap server that is
subscribed to launchpad notifications to use as a source for the event
stream. However there is no preexisting imap server setup anywhere to
use for this. This commit deploys the necessary components for the imap
server. Setting up the launchpad user and the subscription will be a
manual process.
Change-Id: I063ac9da89d227ac59ca1cb8765e3615d7bfdbba
We no longer split physically two different clouds, they are
both located on the same location and VLAN.
Therefore, one inventory and no more region logic.
Change-Id: Id6ed4e50da60f6558fe204ae638a164111fe9a0d
We don't want old, stale copies of our production wiki content
showing up in search engines, so set the mediawiki module parameter
that disallows robots from indexing the site.
Change-Id: If8a2f2c2c00715ecce0ac1aa279f649ec84496a1
Depends-On: Ic62a72555315bd344db338809920a3605f17c8c6
Currently the check is failing because fedora24 isn't referenced
anywhere in manifests/site.pp, so simply copy the appearance of
fedora23 here.
Change-Id: I24e738d5e9ecc01c77970fd1652dd064adb717bb
Thanks to OSIC, we now have access to osic-cloud8. Here we are adding
the required credentials so we can start launching infrastructure on
it.
Change-Id: If278a06c66830e2fd5cb580109eff3884de7a7b8
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
Since it is possible to run more then 1 cloud from OSIC, lets expand
our info to reference osic-cloud1.
Change-Id: I3466779f59a0075516a6f2a584f5f3a111e03826
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
The wiki-dev.openstack.org server will eventually be used to test
newer versions of Mediawiki and extensions/skins. To accommodate
this, also parameterize server backups so that they don't conflict
with production (and are in fact disabled entirely for the dev
site).
Change-Id: I6505d3af87f670e71a440c76873c085d97e5b82f
