3d27e83 | andrewbogott | 20 December 2016, 07:09:02 UTC | Keystone: Move api service to uwsgi/nginx WIP Bug: T150774 Change-Id: I89e9076cd2c830f43b6fabe8ed545d84c425f178 | 22 December 2016, 00:08:28 UTC |
dccbf43 | YuviPanda | 21 December 2016, 21:05:19 UTC | labsdb: Followup to I3156a406f37dd5344273faf5c770c32eddee0e25 Get rid of python-yaml since what we really want is python3-yaml which is already installed too anyway. And this require did not actually make any sense anyway. Change-Id: I138961825ca95ca59dff9b385a360cd8178426d4 | 21 December 2016, 21:08:58 UTC |
d034024 | YuviPanda | 21 December 2016, 20:01:30 UTC | labsdb: Run maintain-dbusers only on active NFS host Change-Id: I3156a406f37dd5344273faf5c770c32eddee0e25 | 21 December 2016, 20:58:21 UTC |
03995fa | Tim Landscheidt | 12 December 2016, 01:02:02 UTC | Tools: Quote arguments in clush Change-Id: I0574ee1b06ea0768a95bb4f5264a4420691234c0 | 21 December 2016, 20:30:49 UTC |
e3fa4d7 | Tim Landscheidt | 12 December 2016, 01:01:32 UTC | Tools: Remove bashisms from clush Change-Id: Ia2da9869a78128fcf155fe6350b8d86524469398 | 21 December 2016, 20:30:33 UTC |
37ec00e | YuviPanda | 15 December 2016, 15:41:10 UTC | labs: Clean out projects that don't exist anymore from mounts I used https://wikitech.wikimedia.org/w/api.php?action=query&list=novaprojects as list. Change-Id: Id2c3575cdd7eb878021018682b6114fc238b4ecb | 21 December 2016, 20:07:25 UTC |
b0dcf92 | YuviPanda | 21 December 2016, 15:54:10 UTC | labsdb: Add delete user functionality to maintain-dbusers Change-Id: I9f4237f00e29f8ee0d828c48c0ac1802b233bf04 | 21 December 2016, 19:55:19 UTC |
68710ac | Filippo Giunchedi | 21 December 2016, 18:46:36 UTC | prometheus: fix syntax for mysql recording rules Change-Id: I837ffc02df437de1f85405b80c87f642d4728acc | 21 December 2016, 19:26:58 UTC |
dc0c642 | andrewbogott | 21 December 2016, 17:29:30 UTC | Pin python-pbr to an old version for Zuul Zuul has a version string that isn't SemVer compliant. Ideally we should rebuild Zuul with a new version string, but for now this will do. Bug: T153877 Change-Id: If7edd5eaed50561b0c0ce57d64f4544eedce8dbd | 21 December 2016, 19:12:31 UTC |
ab7a211 | Daniel Zahn | 20 December 2016, 23:33:35 UTC | install/dhcp: switch all "next-server" from carbon to install1001 Make all networks use install1001 as tftp server where they used to use carbon as tftp server. I have tested it with a single host and installing from install1001 worked just fine now. Bug: T132757 Change-Id: I8ed5efb77b23cde638ec325c1a3fbe95d0cb3c63 | 21 December 2016, 18:46:42 UTC |
9f25820 | Daniel Zahn | 21 December 2016, 18:01:13 UTC | dhcp: re-add europium but in eqiad.wmnet This used to be europium.wikimedia.org but was removed Ica82c78c9aee0493 (T82239). Re-adding it for testing installserver and to install jessie over it. Change-Id: I1bd3af2091f89d5156848b8a12220f7ce7483f08 | 21 December 2016, 18:08:09 UTC |
3514082 | joal | 21 December 2016, 17:40:22 UTC | Correct pivot conf for mediawiki history beta Some settings we used lead to bugs in the UI. Commenting them in that patch Change-Id: I6b5596b11a4b46b5713156354b9e919d3952a4ca | 21 December 2016, 17:40:22 UTC |
e3b81ee | joal | 21 December 2016, 17:05:36 UTC | Update pivot conf for mediawiki history beta Improve last two updates by correcting desription and setting default sort value. Change-Id: Icc157674c9d1fe281a754ee4befa31014a03a6e2 | 21 December 2016, 17:05:36 UTC |
b97347e | milimetric | 21 December 2016, 16:41:30 UTC | Fix typo in pivot config yaml Change-Id: I035913a49f162e10b17f25ef0da608cc22308fe0 | 21 December 2016, 16:54:16 UTC |
bb3aeb4 | Giuseppe Lavagetto | 19 December 2016, 16:41:58 UTC | mediawiki::scaler: check orphaned HHVM threads Due to discrepancies in the timeouts, we might have jobs piling up on HHVM that are long gone in apache2; since this can be a problem (esp. on videoscalers), we start monitorning this. Bug: T153488 Change-Id: Ibcee6af40c678c835c741c100d9fa4df1d99fd86 | 21 December 2016, 16:40:00 UTC |
76af22a | milimetric | 06 December 2016, 16:19:58 UTC | Upgrade edit-history to mediawiki-history-beta The edit_history data cube was an initial run of the mediawiki edit history reconstruction project. We improved the algorithms that refine this data considerably and decided that the "mediawiki history" name was better. This data includes user, page, and revision history, so it goes beyond just editing history. This data cube is also a bit more complicated than the pageview ones in that it includes multiple columns. Operational detail: The edit-history data cube will be deleted. The mediawiki-history-beta data cube will live on until approximately March 2017 when we'll replace it with mediawiki-history. We may update the data a few times until then, but the cube definition should be fairly stable. Change-Id: Ief70b5ed33b57be14a69d4a7f6655a94a9c28acd | 21 December 2016, 16:33:31 UTC |
626b08c | elukey | 21 December 2016, 15:12:10 UTC | Increase Redis Replica Sync retry interval to 2 minutes We usually get false alarms that nobody really pay attention to about Redis replicas not being in sync with their master. The alarms always resolve by themselves in a couple of minutes. Change-Id: I1359ca052bacd59eccb6265713029b4c51d9e9aa | 21 December 2016, 16:29:02 UTC |
3c0588d | Giuseppe Lavagetto | 21 December 2016, 14:39:29 UTC | changeprop: use codfw's restbase in codfw as well This way, dependent updates will be going through codfw as well. Change-Id: I6a09cef5ee07de1a1b0dc0092f7047f8cc565629 | 21 December 2016, 16:01:14 UTC |
9f5ef91 | andrewbogott | 21 December 2016, 00:52:44 UTC | Wikitech: include openstack::clientlib on silver Bug: T153807 Change-Id: I006d1aa5e3c5fbb741ea1c97abc881fcba07e6ef | 21 December 2016, 15:49:42 UTC |
776db4b | Moritz Muehlenhoff | 21 December 2016, 15:01:27 UTC | Add snapshot::testbed to standard snapshot debdeploy group Change-Id: I1069a2cc266c247a101289b1e530f5e87e76ef9b | 21 December 2016, 15:01:27 UTC |
e31e005 | elukey | 21 December 2016, 14:26:35 UTC | Remove MongoDB dependency from statistics cruncher Change-Id: I15a464077e15baba4b15c7c6b3fbebf537229470 | 21 December 2016, 14:27:10 UTC |
7906072 | Moritz Muehlenhoff | 21 December 2016, 08:59:14 UTC | Also follow stat1001 rename in debdeploy grains Change-Id: I54bf76dd0b67782e37f9b5f9fb76e5447face7ad | 21 December 2016, 13:57:40 UTC |
63b2310 | Emanuele Rocca | 20 December 2016, 13:58:46 UTC | varnish cachestats.py: add support for defaults and key_prefix Instead of adding key_prefix in all subclasses, do it once and for all in CacheStatsSender. Also implement support for default values, needed by varnishxcache. Bug: T151643 Change-Id: Ic601f5f10ff29d87aaab90900c6749485818f0ee | 21 December 2016, 13:47:51 UTC |
1f1ba60 | cmjohnson | 19 December 2016, 15:34:46 UTC | Add Eric Evans to analytics-privatedata Bug: T153375 Change-Id: I5899513ebda373d1500cac65d39f2a6071019a43 | 21 December 2016, 10:58:25 UTC |
a79a1ea | Filippo Giunchedi | 19 December 2016, 20:21:55 UTC | package_builder: rebuild Packages only when needed It can take some time and resources to rebuild Packages with many and large packages, skip reconstruction if not needed (e.g. on build failure the list of packages won't change) Change-Id: I218399ce5baf2af76c3167b68e449719e8f41650 | 21 December 2016, 09:41:24 UTC |
489d521 | elukey | 21 December 2016, 09:05:53 UTC | Add hhvm timeouts overrides to hiera for mw116[89] These settings are in the main videoscaler.yaml hiera config but they are not kept in these more specific configurations. Bug: T153488 Change-Id: Idf9c7f0451d12e5089c9e0b5d33719dad7889940 | 21 December 2016, 09:05:53 UTC |
643f301 | elukey | 21 December 2016, 07:54:01 UTC | Increase hhvm threads and transcode capabilities on mw116[89] These videoscalers are power powerful and can manage more transcode jobs at the same time. This change should improve their overall utilization and help a bit reducing the transcode backlog. Bug: T153488 Change-Id: I7e23333ed5cb5e3776b2f5e6f0d025b7e16c3f4a | 21 December 2016, 07:54:01 UTC |
ec910ff | Daniel Zahn | 21 December 2016, 06:28:54 UTC | planet: remove wikimedia.org.au feed The feed doesn't exist anymore, see ticket for the details trying to revive it. More than happy to re-add a working one if there is one in the future. Bug: T133620 Change-Id: I56569282357b31fb940585e7abfcb915fa121adb | 21 December 2016, 06:37:17 UTC |
96d4761 | YuviPanda | 21 December 2016, 05:10:43 UTC | labsdbs: Fixup delete-dbusers - Works without needing to explicitly specify config - Drop users in all dbs Change-Id: Ic30d7198eb073ccb955ebc339c13e9a0bc233b25 | 21 December 2016, 05:10:43 UTC |
61fd808 | YuviPanda | 21 December 2016, 04:54:55 UTC | labs: Fix service unit for maintani-dbusers Change-Id: I64742926b210f973841a9cab06dd5453736618ee | 21 December 2016, 04:55:15 UTC |
c5f0338 | YuviPanda | 14 December 2016, 09:43:07 UTC | labs: maintain-dbusers.py for maintaining labsdb users - Will replace create-dbusers.py - Uses a database in m5 as canonical store of user accounts - Will create different type of accounts for new and old labsdbs Change-Id: I5ced2ca23722267bca57fa234b0a1d6aaa0e9966 | 21 December 2016, 04:48:43 UTC |
fe8151d | Filippo Giunchedi | 20 December 2016, 21:16:50 UTC | prometheus: add aggregation rules for MySQL The rules can be used to speed up dashboards by pre-calculating aggregated metrics. Also collect the metrics into the 'global' instance for a overall view. Change-Id: Id08df46c58f93e71f70ed3a2534658b64dbb047e | 21 December 2016, 01:39:46 UTC |
03fb8dc | Daniel Zahn | 21 December 2016, 01:21:11 UTC | dhcp: switch private-c-eqiad to install1001 as tftp Test installing with install1001 instead of carbon for mw1168. While DHCP is stopped on carbon. After switching one host, now switch one network, finally switch all. Bug: T132757 Change-Id: I4460b587a862694a8a98ee5a9d62236befc13ec4 | 21 December 2016, 01:24:20 UTC |
6d0874b | Dzahn | 20 December 2016, 23:36:56 UTC | Revert "dhcp: set installserver for mw1168 to install1001" This was a test before doing https://gerrit.wikimedia.org/r/#/c/328439/ where we switch it over for all hosts, and it worked fine, so we don't need the special case anymore. This reverts commit 8cc491eaec28fe3773997e4bd5805d8f4e217ff9. Change-Id: I0999b6e343703ba71e5342a38cb5a43d1cd2f357 | 20 December 2016, 23:36:56 UTC |
8cc491e | Daniel Zahn | 20 December 2016, 21:55:33 UTC | dhcp: set installserver for mw1168 to install1001 mw1168 should be reinstalled with a different partman recipe per mail from Luca. And i wanted to test if using install1001 works to replace/reinstall carbon. Change-Id: I8569ab8a5891a3da59b827c581e8a8a36890ec60 | 20 December 2016, 23:08:53 UTC |
5352eb9 | Dzahn | 20 December 2016, 23:05:27 UTC | Revert "install/apt: set install2001 to active host temp." This reverts commit bc6da726521e34e8a1f7ca1fd0a0d86ab5ce15ec. Change-Id: I5c4cd2f612f8bd4c58de40f7ee7373866cf18b6e | 20 December 2016, 23:05:27 UTC |
bc6da72 | Daniel Zahn | 20 December 2016, 23:01:33 UTC | install/apt: set install2001 to active host temp. Have to set install2001 as the active host and let puppet run once, _then_ disable it again to avoid puppet errors with dependency on /etc/acme/challenge-nginx.conf. Change-Id: I064decd0471ea89df89bff026e8d9ed956b66d18 | 20 December 2016, 23:05:12 UTC |
e71d027 | Daniel Zahn | 07 December 2016, 06:17:48 UTC | install: add http & proxy roles on install2001 Make install2001 match install1001 by including the (new) http and proxy classes one subtask of T132757 says. Bug: T132757 Change-Id: Ie1c89f96b7932fc6cfe5a3afc22b93e8dab2e24c | 20 December 2016, 22:52:46 UTC |
59a2284 | Daniel Zahn | 20 December 2016, 22:36:21 UTC | install: add hiera override to skip Letsencrypt cert creation Add a Hiera override to skip the Letsecnrypt cert creation for apt.wikimedia.org on the hosts that currently don't serve it but have the role applied. Bug: T132757 Change-Id: I11a07dbfdc3317422a5f553e9d6b069479d8506c | 20 December 2016, 22:50:39 UTC |
40ec63b | Daniel Zahn | 07 December 2016, 06:06:39 UTC | install: add http & proxy roles on install1001 These 2 roles are the remaining difference between carbon and install1001/2001. This should make them identical. (If everything is puppetized that was on carbon, we will see). These roles have been created by splitting up the existing role::installserver and making sure it always stayed no-op. Bug: T132757 Change-Id: I757cc26a880fb7a82c6eb1d1ca12003fd67a34dc | 20 December 2016, 22:07:39 UTC |
fdd06bc | Filippo Giunchedi | 20 December 2016, 19:14:59 UTC | otrs: use wikimedia_domains from role exim Followup to Iabed0cace1 Change-Id: If541797dcb2e4829d6003632272c4bfe7b0efa74 | 20 December 2016, 20:35:07 UTC |
61e5015 | YuviPanda | 20 December 2016, 20:24:37 UTC | puppetmaster: Do not install default file before common package Otherwise the common package freaks out and refuses to install properly Change-Id: Ie39d101b72e6adb5536304a3f470fce48b98c415 | 20 December 2016, 20:31:57 UTC |
76a3aab | Filippo Giunchedi | 20 December 2016, 20:03:07 UTC | prometheus: introduce prometheus::rule define Also takes care of reloading Prometheus if the rule validates. Change-Id: I67bd5c5cc0b320b1022e7d6180fb5669247f1b36 | 20 December 2016, 20:29:58 UTC |
39e9c80 | YuviPanda | 20 December 2016, 20:04:59 UTC | Revert "puppetmaster: Install self signed CA into system store too" This actually was a red-herring at that point, and might be preventing new standalone puppetmasters from starting up. This reverts commit d48afa3e0bf5781ac1857f9c4c3ad2c7785d84ab. Change-Id: I479a84cec736b60c805f440a95cb37d06754249a | 20 December 2016, 20:05:57 UTC |
76b54e1 | Filippo Giunchedi | 16 December 2016, 23:41:06 UTC | prometheus: add aggregation rules for varnish_exporter Bug: T147424 Change-Id: I5dad7dba00c68011b95ef42c806353e2dd427441 | 20 December 2016, 18:51:30 UTC |
164b790 | Emanuele Rocca | 20 December 2016, 12:01:16 UTC | grafana: Provision Varnish Client Status Code dashboard as JSON This will effectively make the dashboard read only and establish it as operations provisioned dashboard. Change-Id: I7b98b47bccb0c6eb8a1bb9da58eb2568e8ca7152 | 20 December 2016, 18:38:26 UTC |
8b0057c | RobH | 20 December 2016, 17:37:24 UTC | updating sinistra to mwlog2001 renaming the sinistra host to mwlog2001 T153361 Change-Id: Id24332425d54fa9ea3bf2442e07ad7a568502d52 | 20 December 2016, 17:37:24 UTC |
c8a28f1 | elukey | 20 December 2016, 10:16:42 UTC | Move mw116[89] to a single partition recipe The mw116[89] hosts have only one disk and the current recipe creates a very tiny /tmp partition that fills up very quickly (they are videoscalers using the tmp dir for transcode temporary files). The new recipe is basically a single partition without the restriction of the small tmp directory. Bug: T153488 Change-Id: Id336e8addae60645026360b54eeab96bb4a6ff37 | 20 December 2016, 16:27:14 UTC |
50b0949 | Marko Obrovac | 18 December 2016, 20:44:02 UTC | Conftool: Add restbase101[678] and restbase201[012] Bug: T151086 Change-Id: I797bb8a46931687793fb319055be4d17195b56a8 | 20 December 2016, 16:24:32 UTC |
6721fa1 | Marko Obrovac | 20 December 2016, 15:40:11 UTC | ChangeProp: Use RESTBase in the other DC to process async requests Change-Id: I6c79919aa790d267a330cb4c5e3e7008f5bbd325 | 20 December 2016, 15:50:42 UTC |
8669e1e | Marko Obrovac | 20 December 2016, 14:56:40 UTC | Parsoid: provide the full URI in mwapi_server for codfw Change-Id: I5cc045aa6573d19cdc1f4feaa16a289135410d32 | 20 December 2016, 14:56:40 UTC |
1e7493a | Marko Obrovac | 20 December 2016, 12:22:55 UTC | Parsoid: Define the mwapi_server and mwapi_proxy config variables We want Parsoid to be talking to the appservers in eqiad. However, for the codfw->eqiad connection, we are using TLS, for which mwapi_server is used. In eqiad no TLS set up is needed, so use a simple proxy, i.e. mwapi_proxy. Change-Id: Ifeff3620efac7f9a0db4136bc42b6171a8917dd3 | 20 December 2016, 14:17:38 UTC |
79fb1de | Emanuele Rocca | 19 December 2016, 15:31:23 UTC | varnish: remove varnishprocessor The two scripts depending on varnishprocessor, namely varnishrls and varnishmedia, have been ported to cachestats.CacheStatsSender. Remove varnishprocessor. Bug: T151643 Change-Id: Ifd60651d925467dd01ba21bff8f79c5444bb35bb | 20 December 2016, 14:15:30 UTC |
b9521ea | Alexandros Kosiaris | 12 December 2016, 15:19:23 UTC | k8s::apiserver: Amend to support more than labs * Allow using packages, still supporting labs via symlink * Use /etc/default/kube-apiserver * parameterize: kube API port, kubelet_port, service cluster IPs, admission controllers * Have systemd unit notify systemd of startup readiness Change-Id: Ic1d5a1cee782696e1f81a55cc8ba09b2b3059670 | 20 December 2016, 14:08:14 UTC |
5544608 | Alexandros Kosiaris | 12 December 2016, 14:10:28 UTC | k8s::controller: Amend to support more than labs * Allow using packages, still supporting labs via symlink * Use /etc/default/kube-controller-manager * Change the systemd unit name to comply with the rest of the unit naming scheme * Add a cluster-cidr parameter to allow overriding the default Change-Id: I18e0a603193ddbd259bde504bfd29775d9036463 | 20 December 2016, 14:07:48 UTC |
a7dc91a | Alexandros Kosiaris | 12 December 2016, 13:16:50 UTC | kube-scheduler: Amend to support more than labs * Allow using packages, still supporting labs via symlink * Use /etc/default/kube-scheduler Change-Id: Ibec9fa8ad6a8d66ec6bb5f2e27e9cf1a9efacc2d | 20 December 2016, 14:05:04 UTC |
8eb8d2e | Eric Evans | 19 December 2016, 19:59:53 UTC | enable instance restbase1018-c.eqiad.wmnet Bug: T151086 Change-Id: Iccda7d3b86fb6d6b28717324af7fa0786d38f659 | 20 December 2016, 00:28:41 UTC |
d57ed82 | Paladox | 16 December 2016, 16:09:53 UTC | Gerrit: Remove java 7 package We now use java 8 so no use for java 7 now. Change-Id: I22871853828309572a3813f1304edf7d73212059 | 20 December 2016, 00:00:26 UTC |
088d1f7 | Filippo Giunchedi | 19 December 2016, 23:35:22 UTC | Add apache/hhvm exporter to imagescalers Bug: T147423 Bug: T147316 Change-Id: Icd5af8b7d9d7bd6e02d20974099f7f0894d58d63 | 19 December 2016, 23:52:31 UTC |
79fdd15 | Daniel Zahn | 15 December 2016, 22:08:09 UTC | installserver/CI: give shell scripts a file extension For reasons outlined in T148494 we want to give all shell scripts a file extension. This just does that for the ones in installserver module. Bug: T148494 Change-Id: Ic2af0fa32fec2bbb87e61399b7b2421edbed42c1 | 19 December 2016, 23:42:06 UTC |
e488b00 | andrewbogott | 19 December 2016, 23:14:43 UTC | Glance: rename a deprecated setting Change-Id: I9530a369f0afc80e1a7b5493d67122f7521584ee | 19 December 2016, 23:18:53 UTC |
0998931 | Daniel Zahn | 14 December 2016, 01:40:44 UTC | lists/exim: move files from /files to role module Change-Id: Iabed0cace1d5d68d8095695ff04881f6cda557ef | 19 December 2016, 22:57:10 UTC |
8aeacfd | andrewbogott | 19 December 2016, 22:47:20 UTC | Nova: Fix some deprecated settings in nova.conf Change-Id: Ib621ed43c5d8297362bfecd95845213f1341a268 | 19 December 2016, 22:47:20 UTC |
bee59d8 | andrewbogott | 19 December 2016, 22:28:57 UTC | Include some needed __init__ files Change-Id: I10dcc77644688b5e05a9e5396c457010cfe48f6b | 19 December 2016, 22:29:45 UTC |
5e38632 | andrewbogott | 19 December 2016, 22:05:58 UTC | Keystone: Give custom auth plugins entry points Bug: T150773 Change-Id: I0aa5236529fa551d2032f43c0a4ed90ebc72750e | 19 December 2016, 22:18:32 UTC |
aadec36 | Filippo Giunchedi | 19 December 2016, 21:34:30 UTC | admin: fix .bashrc for filippo Change-Id: I877df680d0a686af9c0f698a07a56125419f12b6 | 19 December 2016, 21:34:30 UTC |
e6b2077 | Filippo Giunchedi | 19 December 2016, 20:19:50 UTC | prometheus: sum node_procs_running across clusters Change-Id: Ib315fe7879f11589c4a55311ef70d1c43cf6d5de | 19 December 2016, 21:32:34 UTC |
1d0214e | Andrew Otto | 19 December 2016, 21:10:26 UTC | Fix script path to camus Change-Id: I19d3dc8362ab3d7ff7b147a14602ce6224005ccc | 19 December 2016, 21:10:41 UTC |
cd3ecef | Andrew Otto | 19 December 2016, 21:04:17 UTC | Ensure PYTHONPATH has refinery utils available for analytics cluster camus jobs Change-Id: I2d4db80a7e176926d112141c168dd026353c5564 | 19 December 2016, 21:06:46 UTC |
fe6e5fe | Filippo Giunchedi | 17 December 2016, 03:48:44 UTC | admin: add proxy/on-off for filippo Change-Id: I0ac28da9aaea72bc97d82a64dd4e589a3f6c84e4 | 19 December 2016, 20:28:28 UTC |
4d7a2cc | Daniel Zahn | 16 December 2016, 02:16:00 UTC | contint: fix/move 'backup'-includes, move from node to role Move backup includes from the node level into the role level. They were already duplicated but only of the 2 backup sets in another place. Bug: T150771 Change-Id: Ic1ad70f17893d51590fa50fcb2225622880e40e3 | 19 December 2016, 19:11:50 UTC |
395902c | elukey | 19 December 2016, 15:52:35 UTC | Add the apache/hhvm prometheus exporter to all the mw hosts Bug: T147316 Bug: T147316 Change-Id: I007fbf2138804aaf633295eb788585849c5ae11b | 19 December 2016, 19:07:50 UTC |
587aa75 | Daniel Zahn | 16 December 2016, 02:08:39 UTC | contint: combine contint1001/2001 in a single node regex Now that both hosts are including identical roles we can merge them into one regex. This way we see right away they are identical and we can be sure to avoid subtle differences. We will probably never change this since our actual switch is in Hiera now. Bug: T150771 Change-Id: Icb99b1a801b9168c0d9ba869984d5a8ccf668982 | 19 December 2016, 18:50:57 UTC |
53f8a46 | Brandon Black | 19 December 2016, 18:31:23 UTC | update-ocsp: fixups for Digicert deploy Digicert's responses, which are legitimate, do not include a redundant copy of the signing cert, which makes the response considerably smaller. We need to use our existing on-disk copy of the signing cert in its place for verification, and not look for timestamp windows on the signer (just the OCSP response). They also don't include the nonce in the response, but that doesn't actually create any new issues here. These fixups are also confirmed to work fine with the existing GlobalSign certificates as well. We may want to explore stripping the signing cert from GS for reduced bytes during TLS startup as well, but that can be done separately after validating this all looks ok for clients in general. Change-Id: I91a65da2d3e10c8b5b5b7aa32183a7cb41056dd2 | 19 December 2016, 18:35:59 UTC |
469072c | Marko Obrovac | 16 December 2016, 16:09:00 UTC | Add jdlrobson to the deploy-service group Bug: T153458 Change-Id: Ib3988e333feb1cd1d09b4023c91c7ca855e5c6e6 | 19 December 2016, 18:09:18 UTC |
88c87fb | Marko Obrovac | 16 December 2016, 16:05:35 UTC | Trending Edits: Add the admin group (and add it to SCB) Bug: T153458 Change-Id: I3f21887b3a10b44b57fca8417dd954e70bacda61 | 19 December 2016, 18:05:27 UTC |
8788d72 | Brandon Black | 19 December 2016, 17:40:43 UTC | Add new digicert unified cert files This deploys to cache terminators with OCSP fetching, etc, but does *not* yet configure for user-facing utilization. Change-Id: I3c412bf53f92ce11c513a2f888e5b4ba0cbd6937 | 19 December 2016, 18:04:44 UTC |
97099b9 | elukey | 16 December 2016, 13:34:54 UTC | Add the new user fdans with basic Analytics group permissions Francisco Dans is the new member of the Analytics team and he needs basic Analytics production access credentials to do his work. This code review needs a Ops meeting approval since it grants SUDO permissions. Bug: T153303 Change-Id: I4a58c9028943d908f551137a9c734e89f7f9ac2e | 19 December 2016, 17:56:30 UTC |
1125884 | Brandon Black | 19 December 2016, 17:40:59 UTC | Remove expired unified certs (GS 2015) Change-Id: Ifa0bfbea551728f5c548533ca4d3e7c930d08d85 | 19 December 2016, 17:45:20 UTC |
3de8e11 | Brandon Black | 19 December 2016, 16:53:42 UTC | sslcert: regenerate dhparam.pem It's been several months, re-set the clock JIC. Change-Id: I88334f1fb8b2ef64a88418b1b6d944bd6dc78c7f | 19 December 2016, 17:45:20 UTC |
ed9ca10 | Eric Evans | 19 December 2016, 16:41:13 UTC | enable instance restbase1018-b.eqiad.wmnet Bug: T151086 Change-Id: Icd245358ba63127f576f5c937569108dc90c838b | 19 December 2016, 16:41:13 UTC |
977a2ad | Ariel T. Glenn | 19 December 2016, 16:19:15 UTC | (dumps mirrors) third time's a charm? Change-Id: I31fb5545fd76dbd3943b85d66c3c4a7104b2f425 | 19 December 2016, 16:19:15 UTC |
bfb1077 | Ariel T. Glenn | 19 December 2016, 16:14:54 UTC | (dumps mirrors) aaaand add the missing href close tag. Change-Id: I920ec0d35e183f9b75ac8bf07af3df048dad65b9 | 19 December 2016, 16:14:54 UTC |
0b51ff2 | Ariel T. Glenn | 19 December 2016, 16:09:13 UTC | add Umeå University to public dumps mirrors, yay! Change-Id: I87af9dbd967e5f36618b314aa309d4790364abc1 | 19 December 2016, 16:09:13 UTC |
04f80ff | Ariel T. Glenn | 19 December 2016, 14:30:00 UTC | modify dumps index.html to clarify that cap is only on WMF servers some mirrors pick up this page, and we don't want to give the impression that they impose the same limits on downloaders. Change-Id: Iaf4aab895fbc7931f90ee8b9e847a8b2a9e2c5fc | 19 December 2016, 14:30:00 UTC |
86f0b35 | Giuseppe Lavagetto | 19 December 2016, 13:53:18 UTC | videoscaler: fix and harmonize timeouts Make everything timeout after one day, should be enough for nearly any video. Bug: T153488 Change-Id: I4f7bf27ed038f042ce2a840d20690f037f1adb73 | 19 December 2016, 14:03:55 UTC |
26ae1b9 | Jaime Crespo | 19 December 2016, 13:57:34 UTC | labsdb: Enable view creation on labsdb1009/10/11 Bug: T147052 Change-Id: I11f3d8b80b2ec56a08e5b61a35024a18025226f0 | 19 December 2016, 13:57:34 UTC |
fffc3c1 | Jaime Crespo | 19 December 2016, 10:10:47 UTC | labsdb: Block access to replicas' mysql from almost everywhere Block access to labsdb1009/10/11 from everywhere except: * Monitoring hosts * Admin hosts * Proxy Bug: T147051 Change-Id: I61051f8864dc8a6052c7e105a6d0df38bc51cc23 | 19 December 2016, 13:56:44 UTC |
dced823 | Ariel T. Glenn | 19 December 2016, 13:33:41 UTC | add rest of the dump rsync modules that filter dirs on our side for this adds modules for those mirroring 1,2,3 or 4 last good dump runs Bug: T152954 Change-Id: I04a14ff3a467a2a8d8081c13e317b39d0b435c64 | 19 December 2016, 13:33:41 UTC |
df08a3b | Emanuele Rocca | 16 December 2016, 13:50:38 UTC | varnishrls: port to cachestats.CacheStatsSender Bug: T151643 Change-Id: If791787b9d6901558e9394678c2b08030cda484f | 19 December 2016, 10:44:58 UTC |
8a0901b | elukey | 19 December 2016, 09:18:27 UTC | Repurpose two jobrunners to videoscalers in eqiad Bug: T153488 Change-Id: I11e8b95f52fa2fde0a03c78cdb0472e3da171ad3 | 19 December 2016, 09:53:58 UTC |
ec0f559 | Alexandros Kosiaris | 27 September 2016, 17:19:04 UTC | Rework network::subnets network::subnets was recently made a hiera variable instead of a puppet hash. Various realms were meant to use it as is without really overriding it however. The nature of the different hiera configurations of the realms also meant it was difficult to create a cross realm hiera variable. Due to this, move the data into the module, into a YAML file and away from hiera. Amend network::constants to load it into scope and amend slice_network_constants to use that Puppet variable to get the data instead of using a hiera lookup Change-Id: Ia80e5263db88a0b3a845ae6ac21548725120f0eb | 19 December 2016, 09:14:23 UTC |
6a22a44 | Ariel T. Glenn | 19 December 2016, 08:56:18 UTC | turn off dumping of private tables we have real db backups for these and no one is going to be downloading them so let's not produce them. Change-Id: I2c6dc757aba183f3b6028d7aaf5382ee601cc33a | 19 December 2016, 08:56:59 UTC |
905930a | Ariel T. Glenn | 19 December 2016, 08:51:20 UTC | dumps: fix path to tables yaml file Change-Id: I67f0bf90455c477cef6038e382efedf28759bc8e | 19 December 2016, 08:51:20 UTC |
dd1710f | Ariel T. Glenn | 19 December 2016, 08:43:56 UTC | dumps: fix typo in setup of tables yaml file path Change-Id: I9c7612c527faf698320d5752293b4c896c3b733c | 19 December 2016, 08:44:12 UTC |
63097a7 | Ariel T. Glenn | 08 December 2016, 12:27:49 UTC | remove some dblist paths from dump config settings, no longer needed Change-Id: I4ab6ff3ba024153da4f3fe609320775db2506c41 | 19 December 2016, 08:37:55 UTC |
e8738a8 | Ariel T. Glenn | 08 December 2016, 12:17:37 UTC | move configuration of tables to be dumped out to a yaml file update configuration for dumps accordingly also remove dead 'forcenormal' config setting and commented out settings Change-Id: Idd043f628854200bb04c9c14446708ab3863e20c | 19 December 2016, 08:32:20 UTC |
c0fcab9 | andrewbogott | 15 December 2016, 13:43:01 UTC | Add export OS_INTERFACE=public to observerenv Given semi-modern client packages, this allows us to query keystone without using the (firewalled) admin API endpoint. Bug: T150092 Change-Id: Ia12d9808a3b566e5e54e873c2e74106197dec851 | 16 December 2016, 22:00:23 UTC |
d35cb41 | Eric Evans | 16 December 2016, 21:32:19 UTC | enable instance restbase1018-a.eqiad.wmnet Bug: T151086 Change-Id: I781de0ddae4219f5d82f508eec974a9dd33f64f5 | 16 December 2016, 21:32:33 UTC |
fabe1df | cmjohnson | 16 December 2016, 00:43:03 UTC | Adding prometheus1003-4 to netboot.cfg file T152504 Change-Id: I320ea26edaf1160074471272abd698dc80c9e022 | 16 December 2016, 19:27:26 UTC |