Bring sto2 back online and monitor.
This reverts commit 51d6d25097be440f1e06512e57e879afde52f91a.
Change-Id: Id4df6e5605dc61035fb14575744dfdd02cb66cf3
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
We skip long-running dsvm jobs for nova changes that are
only in the api-ref tree, and we should do the same for
the placement-api-ref tree, like in this change:
Id89d7c16ca53938e2bc18e904e8d13477ceb15f7
Change-Id: Ife1e1ad0fa2dd54f5f92d72e044e11b6987ba15e
Mirror replacemement in this region is complete, so start using it
again.
This reverts commit ce12a892ef2acc557d33e86138c221a651e87b07.
Change-Id: I1dc1669520bdb3c20dcdd8186f83b4dbb1fdcf27
With infracloud reduced by 50%, enable upload again so we can monitor
network bandwidth.
This reverts commit 2a384ab5720711357893b70bd6dd039e4fd0ebe6.
Change-Id: I5ad81f4936b97705ef45d8d077071c252840b519
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
According to discussion in the mailing list [0] using UUID tokens
is currently resulting in errant Keystone-related bugs. Further,
UUID tokens are currently deprecated and will be removed during
R release [0]. So we should use Fernet tokens instead.
[0] http://lists.openstack.org/pipermail/openstack-dev/2017-August/121067.html
Change-Id: I34d324575c8117022724a944c034f089eb0c7541
Related-Bug: #1710237
In preparation for a mirror server replacement, temporarily cease
using rax-ord. This gives us an opportunity to allow htcacheclean to
complete on the proxy cache volume, allows us leeway to move the
Cinder volumes from the old mirror to its replacement, and also
stops the recurrence of pip hash mismatch errors in jobs while we
get the new server online. We're upgrading from 4GB to 8GB
performance flavor because we seem to be spending a fair amount of
time at or over the bandwidth limit for the current flavor.
Change-Id: I954fda2fecef0832e81511d07c8526036d2dabf9
Zuul v3 is running the docs jobs on check and gate, but not the
publishing job in post yes. Remove use of the template and retain
only the post job.
Change-Id: Ief2feaab07527ce1bd93fd4ec8e983ac3b1b8dd0
gate-grenade-dsvm-networking-midonet-v2-ubuntu-xenial examines
an upgrade scenario from v2 monolithic plugin to ml2.
Because v2 monolithic plugin has been removed in Pike cycle,
This job will stop working on master once branches for grenade
jobs are updated for Pike. [1]
[1] Ifb7f7c336523f86fd7a2c18def2ddbbbf656f2fb
Closes-Bug: #1710236
Change-Id: I3a8b6444b79ab1e4088cd1a8d6e7ae51451db87c
- Add templates for Pike translation
- Set up horizon and plugin projects who currently have
stable/pike branch listed in https://translate.openstack.org/
Change-Id: I224ffceb49fad380ab894dc2d8b476b8edf76aab
We've been seeing a large number of jobs timeout in this region.
Looking at logstash, it appears we may have some misbehaving compute
nodes our VMs run on. Both nodes and mirror have been experiencing
bad IO issues.
Change-Id: I413768a15af40e9c4d9c594258d03bd1cbf7988f
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
Now that we have a working reverse proxy cache for api.rubygems.org,
add the .gemrc file and start using it.
Change-Id: I6d8cd1d1353fc28469aceb16d487cd9b39eda73b
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
Step one in renaming this is to make a new job so we can consume it in
openstack-zuul-jobs. Then we can remove the old one.
Change-Id: Iea819609d2f4e343c554a08392fa0620841c3dac