We use extras is enabled by default, so we'll need to mirror it too.
Change-Id: Ib33aa49d58d2a71062dd82bbcdbffd736e221590
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
Now that we have done our initial vos release of mirror.centos, we can
enable it again on mirror-update.o.o.
Change-Id: I71f095e7a0949f5d9c7d18a1e64b9b28d70b6cae
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
The hideci script previously tried to look up the pipeline of a CI
comment by parsing the typical format (from zuul). It would then
group on this pipeline name and consider multiple comments in one
pipeline to be rechecks.
This doesn't work when two different CI's both use the same pipeline
name. Only the latest CI would be shown.
Instead consider each CI+pipeline to be unique.
This change calls each CI in the results table by "CI Name + Pipeline".
Previously if no pipeline was supplied (eg if not using zuul),
"CI Name + 'check'" would be used. Now just "CI Name" will be displayed.
Change-Id: Ie0f90efeba95ab510dacfb6a0cbe437e9ad2afe0
There are a couple of edge cases with the recent PEP503 workaround
where packages with both .'s and -'s in the name are unable to be
translated to their old 'safe names'.
Add specific workarounds for known packages.
Change-Id: I72bbdd4af7200f8565d7c6b7fa6d3f268c783b24
Temporary work around so our mirrors work with the latest pip release
that expects .'s to be replaced with -'s.
This is crude but considered easier than pinning pip across our services.
It may miss some edge cases so we'll see how it goes.
This should be reverted once bandersnatch supports pep503:
https://bitbucket.org/pypa/bandersnatch/pull-requests/20/fully-implement-pep-503-normalization/diff
Change-Id: Ic4acf4c1d4e78d2524ea49356c61916947fd9a3d
This reverts commit 4e5a0dc4586d313b8692e8f9600ab7be640590da.
The read only volumes are happy again.
Change-Id: I78dca90b103153cc8a7dd21b6df99bb3875d41a7
We needed to update our rsync flags because we are seeing issue with
chown / chgrp. Additionally, hard-linking is problematic.
Add missing k5start for mkdir.
Update crontab with missing mirror name.
Change-Id: I9addeae6a6ccf86684d3d5ebd42d4f8d96a9b916
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
To avoid an unfortunate bootstrapping race, make sure we always run
`apt-get update` after replacing the sources list but before any
packages are installed.
Change-Id: I30b9cb3d6cbfa5df59dd1c1eef88bf486f96a52c
This is the first step to getting an AFS mirror for centos going. We
still need to add RPM validation and enabled automatic vos release.
However, this will atleast enabled us to do the initial import of the
RPMs into AFS which will take a few days to mirror properly.
Change-Id: Ic3b0ca5f5c9321b7e5048f4000ca58266d117083
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
While it's hopefully not useful for much longer, this is the script that
safely cleans FIPs from nodepool clouds that use neutron FIPs.
Change-Id: I6f6f2c5f119a40060f46ea1f244a65601f8b7bb9
The new regex shows all jobs - but it does not show anymore Success/fail and the job run times.
This reverts commit 48e8f7e93739bf6ce50e9087c8085a80c76dcdf1.
Change-Id: I99896753342a9082f922df645ece82c5a8af78e7
The vos release on AFS for PyPI is in a bad place emotionally right now.
So as not to stress it out too much more, stop vos releasing until we
can get it fixed. Also, point the AFS client mirrors to the read/write
path instead of the read only path so that the gate can get PyPI updates
again.
Change-Id: I5d9ac9cc56c16dacad22c1da157cad7f5480c301
There have been a couple bug fix release between latest 1.7.x and 1.7.3
which we are running. Pull in this fixes in hopes that it makes shard
allocation more reliable.
Change-Id: I48f46663c73cb178ca3cf95a166b3df2ea121459
This regex dates from the original commit
(Idb4ad8e6079165d681271987a92cab5d8b7c81be) where I think the negative
matches are intended to stop matching between groups.
I don't think this is necessary. In particular, the negative match on
the last group ($4) means that it will not match if there is a URL in
the free-form text portion.
I've simplified this and also added some comments to make it clear
what it is looking for.
Change-Id: I863d88e9a7ed2fd41924b8fc4a12dbea3ee2b205
As a host may not have wget add logic to try curl too.
Change-Id: I9a2ffeff08926eec5f4fee04fde0ddcc19004493
Signed-off-by: Paul Belanger <pabelanger@redhat.com>
openstack-vitrage IRC channel is used by Vitrage project for IRC communication. We would like to log this channel.
Change-Id: Ifcb62a0d1a5474817028a7418b0431d32a5808bf
We can readd them when they return
Running ansible in ad-hoc mode against 'all' hosts causes many errors
because these hosts don't exist any more.
Change-Id: If095df0811ca6d2fc04542044958055b9b835170
In addition to enabling root logins with host matches in sshd config, we
need to put localhost into the key options line.
Change-Id: I09c4120f152fd69b54cad259394721d208efdb23
We can only get the volume attach device if we are attaching a volume.
Check if the volume is being attached and only determine the attachment
location in that case to avoid errors.
Story: 2000569
Change-Id: I4adc5e23abdfc0627a0850f845e2333d3bd25e63
I think that the puppetdb module expects the package to do this work,
and that's why it doesn't do this for us. There may also be an issue
with the puppetdb module setting up a config file under /etc and the
puppetdb application looking for a config file under /opt/puppetlabs
Change-Id: Ide2b3a5841ac07fd7e154e8d26ce0b036f734a41
